surgery for lumbar radiculopathy/sciatica · surgery for lumbar radiculopathy/sciatica draft...

Surgery for Lumbar

Radiculopathy/Sciatica

Draft evidence report

February 19, 2018

Health Technology Assessment Program (HTA)

Washington State Health Care Authority

PO Box 42712 Olympia, WA 98504-2712

(360) 725-5126 www.hca.wa.gov/hta

[email protected]

http://www.hca.wa.gov/hta

Prepared by:

RTI International–Evidence-based Practice Center

Research Triangle Park, NC 27709

Lead Investigator:

Leila Kahwati, MD MPH

This health technology assessment is based on research conducted by RTI International under

contract to the Washington Health Care Authority (HCA) (Contract No. K1959). The findings

and conclusions in this document are those of the authors, who are responsible for its contents;

the findings and conclusions do not necessarily represent the views of the Washington HCA.

Therefore, no statement in this report should be construed as an official position of Washington

HCA.

The information in this report is intended to help the Washington HCA make well-informed

coverage determinations and thereby improve the quality of health care services. This report is

not intended to be a substitute for the application of clinical judgment. Anyone who makes

decisions concerning the provision of clinical care should consider this report in the same way as

any medical reference and in conjunction with all other pertinent information (i.e., in the context

of available resources and circumstances presented by individual patients).

This document is in the public domain and may be used and reprinted without permission except

those copyrighted materials that are clearly noted in the document. Further reproduction of those

copyrighted materials is prohibited without the specific permission of copyright holders.

None of the investigators has any affiliations or financial involvement that conflicts with the

material presented in this report.

Acknowledgments

The following individuals contributed to this report:

Co-Investigator: Rachel Palmieri Weber, PhD

Clinical Advisor: Moe Lim, MD

Analysts: Rachel Clark, BA; Katrina Burson, RN, MS

Scientific Reviewer: Meera Viswanathan, PhD

Library/Document Preparation: Mark Howell, MLS; Loraine Monroe; Laura Small, BA

WA – Health Technology Assessment February 19, 2018

Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page i

Contents

Executive Summary ................................................................................................................ ES-1 Structured Abstract .......................................................................................................... ES-1 ES-1. Background ........................................................................................................... ES-3

ES-2. Methods ................................................................................................................. ES-4 ES-3. Results ................................................................................................................... ES-8 ES-4. Discussion ........................................................................................................... ES-29 ES-5. Conclusion ........................................................................................................... ES-32

Full Technical Report ................................................................................................................... 1 Structured Abstract ................................................................................................................ 1 1. Background ...................................................................................................................... 1

1.1 Purpose ........................................................................................................................ 1 1.2 Condition Description ................................................................................................. 1 1.3 Disease Burden ............................................................................................................ 1 1.4 Technology Description .............................................................................................. 2 1.5 Regulatory Status ........................................................................................................ 3

1.6 Policy Context ............................................................................................................. 4 1.7 Washington State Agency Utilization Data ................................................................. 4

2. Methods............................................................................................................................ 4 2.1 Research Questions and Analytic Framework for Systematic Review of Primary

Research Studies .......................................................................................................... 4

2.2 Clinical Practice Guideline Synthesis ....................................................................... 10 3. Results ............................................................................................................................ 10

3.1 Literature Search ....................................................................................................... 10 3.2 Efficacy ..................................................................................................................... 12

3.3 Safety ......................................................................................................................... 66 3.4 Cost and Cost-Effectiveness ...................................................................................... 80

3.5 Clinical Practice Guideline Synthesis ....................................................................... 87 4. Discussion ...................................................................................................................... 96

4.1 Summary of the Evidence ......................................................................................... 96

4.2 Limitations of the Evidence Base ............................................................................ 102 4.3 Other related HTAs ................................................................................................. 104 4.4 Selected payer coverage policies ............................................................................. 104

4.5 Limitations of this HTA .......................................................................................... 110 4.6 Ongoing Research and Future Research Needs ...................................................... 110

5. Conclusion ................................................................................................................... 110 6. References .................................................................................................................... 112

List of Appendices

Appendix A. Search Strategy ...................................................................................................... A-1 Appendix B. Additional Methods ............................................................................................... B-1 Appendix C. Evidence Tables ..................................................................................................... C-1 Appendix D. Excluded Studies ................................................................................................... D-1

Appendix E. Individual Study Risk of Bias Assessments ........................................................... E-1 Appendix F. Meta-analyses ......................................................................................................... F-1


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page ii

List of Figures

Figure ES-1. Analytic framework for HTA on surgery for lumbar radiculopathy ................ ES-5

Figure 1. Analytic framework for HTA on surgery for lumbar radiculopathy ...................... 5

Figure 2. Study flow diagram for HTA on surgery for lumbar radiculopathy ..................... 11

Figure 3. Evidence map of surgery compared with nonsurgical interventions for treatment

of symptomatic lumbar radiculopathy .................................................................. 97

Figure 4. Evidence map of minimally invasive surgery compared with discectomy or

microdiscectomy for treatment of symptomatic lumbar radiculopathy ................ 98

Figure 5. Evidence map of microdiscectomy compared with discectomy for treatment of

symptomatic lumbar radiculopathy ....................................................................... 99

Figure 6. Evidence map of repeat lumbosacral decompression compared with spinal cord

stimulation for treatment of recurrent symptomatic lumbar radiculopathy ........ 100

Figure 7. Evidence map of revision endoscopic discectomy compared with revision

microdiscectomy for treatment of recurrent symptomatic lumbar radiculopathy

............................................................................................................................. 101

List of Tables

Table ES-1. Surgical interventions used to treat lumbar radiculopathy ............................... ES-4

Table ES-2. Population, intervention, comparator, outcome, timing, setting and other study

selection criteria for HTA on surgery for lumbar radiculopathy ...................... ES-6

Table ES-3. Surgical and comparator interventions used among 22 included studies

for EQ1.............................................................................................................. ES-9

Table ES-4. Summary of efficacy outcome findings and strength of evidence ratings

comparing surgery to nonsurgical interventions in persons with symptomatic

lumbar radiculopathy (EQ1) ........................................................................... ES-10


comparing minimally-invasive surgery to standard surgery in persons with

symptomatic lumbar radiculopathy (EQ1) ...................................................... ES-13


comparing microdiscectomy to discectomy in persons with symptomatic

lumbar radiculopathy (EQ1) ........................................................................... ES-15

Table ES-7. Summary of efficacy findings and strength of evidence ratings comparing

repeat lumbosacral decompression surgery with spinal cord stimulation for

treatment of lumbar radiculopathy relapses (EQ2) ......................................... ES-17

Table ES-8. Summary of efficacy findings and strength of evidence ratings comparing

revision endoscopic discectomy with revision microdiscectomy for

treatment of relapsed lumbar radiculopathy (EQ2) ........................................ ES-18

Table ES-9. Summary of safety outcome findings and strength of evidence ratings

comparing surgery to nonsurgical interventions in persons with symptomatic

lumbar radiculopathy (SQ1) ........................................................................... ES-20


comparing minimally-invasive surgery to standard surgery in persons with

symptomatic lumbar radiculopathy (SQ1) ...................................................... ES-21


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page iii


comparing microdiscectomy to discectomy in persons with symptomatic lumbar

radiculopathy (SQ1) ........................................................................................ ES-23


comparing revision surgery to spinal cord stimulation or microdiscectomy

in persons with relapsed lumbar radiculopathy (SQ1) ................................... ES-24

Table ES-13. Summary of cost-effective findings comparing surgery to nonsurgical

interventions in persons with symptomatic lumbar radiculopathy (CQ1) ...... ES-25

Table ES-14. Summary of cost-effective findings comparing minimally-invasive

surgery to standard surgery in persons with symptomatic lumbar

radiculopathy (CQ1) ....................................................................................... ES-25

Table ES-15. Summary of costs comparing microdiscectomy to discectomy in persons

with symptomatic lumbar radiculopathy (CQ1) ............................................. ES-26

Table ES-16. Synopsis of clinical practice guidelines related to lumbar radiculopathy or

herniated intervertebral lumbar disc ............................................................... ES-26

Table ES-17. Overview of payer coverage policies .............................................................. ES-30

Table 1. Description of surgical interventions used to treat lumbar radiculopathy .............. 2

Table 2. Population, intervention, comparator, outcome, timing, setting and other study

selection criteria for HTA on surgery for lumbar radiculopathy ............................ 6

Table 3. Study and population characteristics of the seven randomized controlled trials

comparing surgery to nonsurgical interventions for management of lumbar

radiculopathy (EQ1).............................................................................................. 13

Table 4. Study and population characteristics of the 15 randomized controlled trials

comparing alternative surgical interventions for management of lumbar


Table 5. Surgical and comparator interventions used among 22 included studies for

EQ1 ....................................................................................................................... 22

Table 6. Summary of efficacy outcomes reported by included studies, including

score range, minimally important clinical difference, and required sample

size to detect various between-group differences ................................................. 24

Table 7. Summary of findings and strength of evidence ratings comparing surgery to

nonsurgical interventions for pain in persons with symptomatic lumbar


Table 8. Summary of findings and strength of evidence ratings comparing minimally

invasive surgery to standard surgery for pain in persons with symptomatic

lumbar radiculopathy (EQ1) ................................................................................. 32

Table 9. Summary of findings and strength of evidence ratings comparing

microdiscectomy to discectomy for pain in persons with symptomatic lumbar

radiculopathy......................................................................................................... 38


nonsurgical interventions for functioning/disability in persons with

symptomatic lumbar radiculopathy (EQ1) ............................................................ 40


invasive surgery to standard surgery for functioning/disability in persons with



Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page iv


microdiscectomy to discectomy for functioning/disability in persons with

symptomatic lumbar radiculopathy ....................................................................... 49


nonsurgical interventions for quality of life in persons with symptomatic



invasive surgery to standard surgery for quality of life in persons with



microdiscectomy to discectomy for quality of life in persons with symptomatic



nonsurgical interventions for neurologic symptoms in persons with



invasive surgery to standard surgery for neurologic symptoms in persons with



nonsurgical interventions for outcomes related to return to work in persons

with symptomatic lumbar radiculopathy (EQ1) ................................................... 55


invasive surgery to standard surgery for return to work outcomes in persons

with symptomatic lumbar radiculopathy (EQ1) ................................................... 57

Table 20 Summary of findings and strength of evidence ratings comparing

microdiscectomy to discectomy for return to work outcomes in persons with


Table 21. Study and population characteristics of the two randomized controlled trials

comparing revision surgical interventions to spinal cord stimulation or an

alternative revision surgery for the management of lumbar radiculopathy

relapses (EQ2) ....................................................................................................... 62

Table 22. Summary of findings for pain, functioning, neurological symptoms and

quality of life for RCTs for repeat lumbosacral decompression surgery

compared with spinal cord stimulation for treatment of lumbar radiculopathy

relapses (EQ2) ....................................................................................................... 63

Table 23. Summary of findings for pain, functioning, quality of life, neurologic

symptoms and return to work comparing revision endoscopic discectomy to

revision microdiscectomy for treatment of lumbar radiculopathy relapses (EQ2) 64


nonsurgical interventions for mortality in persons with symptomatic lumbar

radiculopathy (SQ1) .............................................................................................. 67


invasive surgery to standard surgery for mortality in persons with s

ymptomatic lumbar radiculopathy (SQ1) ............................................................. 68


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page v


microdiscectomy to discectomy for mortality in persons with symptomatic

lumbar radiculopathy (SQ1) ................................................................................. 69

Table 27. Summary of findings and strength of evidence ratings of revision surgery for

mortality in persons with recurrent lumbar radiculopathy (SQ1) ......................... 69

Table 28. Summary of findings and strength of evidence ratings for surgical morbidity

in persons with symptomatic lumbar radiculopathy who undergo surgical

intervention (SQ1) ................................................................................................ 70


invasive surgery to standard surgery for surgical morbidity in persons with

symptomatic lumbar radiculopathy (SQ1) ............................................................ 71


microdiscectomy to discectomy for surgical morbidity in persons with



surgical morbidity in persons with recurrent lumbar radiculopathy (SQ1) .......... 74


nonsurgical interventions for reoperations in persons with symptomatic



invasive surgery to standard surgery for reoperations in persons with



microdiscectomy to discectomy for reoperations in persons with symptomatic



reoperations in persons with recurrent lumbar radiculopathy (SQ1) .................... 78


nonsurgical interventions for persistent opioid use in persons with



invasive surgery to standard surgery for persistent opioid use in persons with



persistent opioid use in persons with recurrent lumbar radiculopathy (SQ1) ....... 80

Table 39. Study characteristics of the seven studies that evaluated cost effectiveness of

surgery for lumbar radiculopathy (CQ1) .............................................................. 81

Table 39. Study characteristics of the seven studies that evaluated cost effectiveness of

surgery for lumbar radiculopathy (CQ1) (continued) ........................................... 82


nonsurgical interventions for cost and cost-effectiveness in persons with

symptomatic lumbar radiculopathy (CQ1) ........................................................... 83


microdiscectomy to discectomy for direct surgical costs in persons with

symptomatic lumbar radiculopathy (CQ1) ........................................................... 87


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page vi

Table 43. Clinical practice guidelines related to lumbar radiculopathy or herniated

intervertebral lumbar disc ..................................................................................... 88

Table 44. Overview of payer coverage policies .................................................................. 104

Table 45. Selected payer coverage for surgery for lumbar radiculopathy .......................... 105


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page vii

List of Abbreviations

AMD Absolute mean difference

ARD Absolute risk difference

CI Confidence interval

HTA Health technology assessment

NR Not reported

NS Not significant

RCT Randomized controlled trial

RR Relative risk ratio

U.K. United Kingdom

U.S. United States


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page ES-1

Executive Summary

Structured Abstract

Purpose: To conduct a health technology assessment (HTA) on the efficacy, safety, and cost of

surgery for the treatment of symptomatic lumbar radiculopathy, also referred to as sciatica.

Data Sources: PubMed from January 2007 through November 9, 2017; clinical trial registry;

government, payor, and clinical specialty organization websites; bibliographies of relevant

clinical practice guidelines and systematic reviews.

Study Selection: Using a priori criteria, we selected English-language primary research studies

published in any year that were conducted in highly developed countries that enrolled adults with

symptomatic lumbar radiculopathy and compared surgery for radiculopathy to nonsurgical

interventions or that compared alternative surgical procedures. We selected trials or cost analyses

that reported efficacy outcomes (pain, functioning and disability, quality of life, neurological

symptoms, return to work) safety (mortality, surgical morbidity, reoperations, persistent opioid

use), or cost per quality-adjusted life year. We also selected the most recently published relevant

clinical practice guidelines (CPG) for quality appraisal and synthesis.

Data Extraction: One research team member extracted data and a second checked for accuracy.

Two investigators independently assessed risk of bias of included primary research studies and

conducted a quality assessment of included CPGs.

Data Synthesis: We included 25 primary research studies published between 1983 and 2017.

Twenty-four randomized controlled trials (RCTs) provided findings related to efficacy and

safety, 7 cost analyses provided findings related to cost-effectiveness. One RCT was rated as low

risk of bias, 12 were rated as having some concerns for bias, and 12 were rated as high risk of

bias. We identified 14 clinical practice guidelines.

Microdiscectomy and discectomy reduced leg pain by 6 to 26 points more than non-surgical

interventions as measured on a 0 to 100 point visual analogue scale of patient-reported pain at up

to 26 weeks followup; these findings did not persist at followup 1 year or later. Similar findings

were observed for functioning and disability as measured by the Oswestry Disability Index,

Roland-Morris Disability Questionnaire, and SF-36 Physical Functioning subscale. Surgery and

non-surgical interventions produced similar improvements in quality of life, neurologic

symptoms, and return to work. No surgical deaths occurred in any study and surgical morbidity

was infrequent. The incidence of reoperations among participants who underwent surgery ranged

from 0% to 10%; persistent opioid use was similar between groups. Studies reported higher

quality-adjusted life years for participants who underwent surgery compared to non-surgical

interventions, but similar or higher costs. The average cost per quality-adjusted life year gained

from a health care payor perspective ranged from $51,156 to $83,322 in 2010 US dollars.

Compared to microdiscectomy or discectomy, minimally-invasive surgery produced similar

improvements in pain, function/disability, quality of life, and neurologic symptoms, but resulted

in return to work 4 to 15 weeks sooner. No surgical deaths occurred in any studies and with few



exceptions, surgical morbidity was similar. The incidence of reoperations across study groups

ranged from 2% to 64.5%; 2 of the 10 relevant studies reported much higher incidence of

reoperation among participants who underwent minimally-invasive procedures (64.5% vs. 2.5%

in 1 RCT; 44% vs. 16% in the other RCT) but the other 8 RCTs reported a similar incidence.

Compared to discectomy, microdiscectomy results in similar improvements pain, function and

other efficacy outcomes, and similar incidence of surgical morbidity and reoperations.

The four higher quality clinical practice guidelines we identified were in general agreement

about recommending discectomy or microdiscectomy (and related decompressive procedures) as

acceptable treatment based on evidence that it improves outcomes in the short to medium term.

Limitations: The included RCTs were limited by methodologic designs that increased risk for

bias, including extensive participant crossover, lack participant and outcome assessor blinding,

and inadequate randomization and allocation concealment in some studies. Many RCTs were

underpowered for outcomes of interest for this review, leading to imprecision for many effect

estimates reported. This HTA was limited to English-language studies reporting intent-to-treat

analyses; it did not include observational studies or ‘as-treated’ analyses reported by some RCTs.

Conclusions: Compared with non-surgical interventions, surgery reduces pain and improves

function more up to 26 weeks followup, but this difference does not persist at 1 year or longer.

Minimally invasive surgery, microdiscectomy, and discectomy are generally comparable with

respect to efficacy and surgical morbidity; findings were mixed for reoperations. Surgery may be

cost-effective when compared with non-surgical interventions, depending on a decision maker’s

willingness to pay threshold, but the evidence is inconclusive about the cost-effectiveness of

minimally invasive surgery. Nearly all findings are based on evidence with low to very low

certainty.



ES-1. Background

We designed this health technology assessment (HTA) to assist the State of Washington’s Health

Technology Clinical Committee with determining coverage for selected surgical interventions to

treat symptomatic lumbar radiculopathy, also known as sciatica.

ES-1.1 Condition Description

Lumbar radiculopathy is a clinical syndrome characterized by radiating leg pain, with or without

motor weakness, and sensory disturbances in a myotomal or dermatomal distribution. Lumbar

radiculopathy is a heterogenous condition that may present acutely (as in the case of an acute

disc herniation) or more insidiously (as in the case of spondylosis). The objective of treatment

for radiculopathy is symptom relief through nonsurgical management of symptoms, or surgical

intervention to address the underlying causative mechanism, or both.

ES-1.2 Disease Burden

Estimates of the prevalence of lumbar radiculopathy vary widely, likely due to variation in

definitions and differences between self-reported and clinically assessed symptoms.1 Several

studies report the lifetime prevalence of lumbar radiculopathy ranges from approximately 3% to

43%.1-3 Previous history of axial low back pain is an established risk factor for radiculopathy.4,5

Other risk factors include a prior history of trauma, prolonged driving, pregnancy, and jobs

requiring manual labor.2,4,6 Several studies have shown a genetic linkage for spinal canal size as

well as disc herniation.4,7

ES-1.3 Technology Description

Table ES-1 lists the surgical procedures used to treat lumbar radiculopathy. Standard, open

surgical interventions remove parts of the intervertebral disc, with or without additional

decompression of spinal nerve root(s) through removal of parts of the bony vertebrae, facet joints

(e.g., laminectomy or partial facetectomy) and/or other soft tissues impinging on the nerve

root(s). Decompression and disc removal interventions are often performed with a microscope or

other magnifying instrument (“micro” approaches). Such an approach makes it possible to

minimize the length of incision and area of dissection, thereby reducing the degree of structural

alteration to surrounding tissues. Both standard open and “micro” approaches allow for direct

visualization of the disc and surrounding structures. When spondylolisthesis is present

preoperatively, or when spinal instability resulting from decompressive or disc removal

procedures is likely (e.g., the surgery involves vertebrae at higher lumbar levels or the nerve root

compression is in the cephalocaudal direction), then arthrodesis, (i.e., spinal fusion) may also be

required. In contrast, minimally invasive surgical interventions use either an endoscopic

approach to allow direct visualization of the surgical field and anatomy, or use a percutaneous

approach that does not allow direct visualization of the disc and surrounding tissue. Further,

these procedures use mechanical (manual or automated), radiofrequency thermal, coblation (also

known as plasma), or laser-assisted techniques for disc removal, destruction, or decompression.



Table ES-1. Surgical interventions used to treat lumbar radiculopathy

Disc removal procedures

Discectomy

Microdiscectomy

Decompression procedures

Laminectomy

Microlaminectomy

Laminotomy

Foraminotomy

Minimally invasive procedures

Percutaneous and/or endoscopic discectomy, discoplasty, nucleotomy, or nucleoplasty

ES-1.4 Regulatory Status

Surgical procedures are not regulated; however, the U.S. Food and Drug Administration (FDA)

(through the 510(k) process) may regulate surgical instruments and devices used in surgery. The

FDA has approved several devices for cutting, grinding, and aspiration of disc material during

discectomy and for ablation and coagulation. FDA has also approved laser instruments for

incision, excision, resection, ablation, vaporization, and coagulation of tissue during surgical

procedures including but not limited to discectomy. See the Full Report for a detailed

description.

ES-1.5 Policy Context

Numerous surgical and nonsurgical approaches to the management of lumbar radiculopathy are

routinely used within current clinical practice. In addition to standard surgical techniques (e.g.,

discectomy with or without laminectomy), minimally invasive surgical techniques that use

percutaneous, endoscopic, or laser-assisted approaches are now available. The State of

Washington Health Care Authority selected surgery for lumbar radiculopathy as a topic for an

HTA based on medium concerns for efficacy, medium concerns for safety, and high concerns for

cost.

ES-2. Methods

This HTA includes two separate, but related components. The first component is a systematic

review of primary research studies and the second component is a quality appraisal and synthesis

of relevant clinical practice guidelines.

ES-2.1 Research Questions and Analytic Framework for Systematic Review

of Primary Research Studies

We developed the following research questions and analytic framework (Figure ES-1) to guide

the systematic review of primary research studies:

Efficacy Question 1 (EQ1). In adults with symptomatic lumbar radiculopathy, what is the

effectiveness and comparative effectiveness of surgical interventions?



Efficacy Question 2 (EQ2). In adults with symptomatic lumbar radiculopathy, does

effectiveness or comparative effectiveness of surgical interventions vary for patients who are not

employed because of disability or patients who are undergoing recurrent surgery for relapse?

Safety Question 1 (SQ1). In adults with symptomatic lumbar radiculopathy, what are the

adverse events associated with surgical interventions?

Cost Question 1 (CQ1). In adults with symptomatic lumbar radiculopathy, what is the cost-

effectiveness of surgical interventions?

Figure ES-1. Analytic framework for HTA on surgery for lumbar radiculopathy

Adults with

symptomatic

lumbar

radiculopathy

Surgical

interventions to

reduce pain,

symptoms, and

improve function

Pain

Neurological symptoms

Health-related quality of life

Physical, psychological, and

social functioning

Return to work

Reoperations for relapse/

recurrent symptoms

$/quality-adjusted life year

gained

$/disability-adjusted life

year/gained

Surgery-related morbidity and mortality

Reoperations for complications

Persistent opioid use

EQ1

SQ1

CQ1

EQ2

Abbreviations: CQ=cost question; DD=developmental disability; EQ=efficacy question; SQ=safety question

ES-2.1.1 Data Sources and Search

The full search strategy is detailed in Appendix A of the Full Report. We searched MEDLINE®

(via PubMed) from 2007, the Cochrane Library, a clinical trials registry (clinicaltrials.gov) and

relevant government, payer, and health care professional society websites for relevant English-

language studies. In addition, we reviewed the reference lists of relevant studies, practice

guidelines, and other HTAs to identify any relevant articles not found through the electronic

search and to identify studies published prior to 2007. We used medical subject headings (MeSH

terms) and text words associated with the surgical interventions of interest combined with MeSH

terms for radiculopathy and lumbar disc disease.

ES-2.1.2 Study Selection

Table ES-2 summarizes the study selection criteria related to the population, intervention,

comparator, outcomes, time period, and setting that defined the scope of this HTA. We screened

titles and abstracts and full-text articles based on these study selection criteria.



Table ES-2. Population, intervention, comparator, outcome, timing, setting and other study selection criteria for HTA on surgery for lumbar radiculopathy

Domain Included Excluded

Population Adults age 18 years and over with symptomatic lumbar radiculopathy (i.e., sciatica) unrelated to infection, cancer, inflammatory, congenital, or traumatic etiologies. For studies of mixed populations, results must be stratified and reported separately for patients with lumbar radiculopathy.

Adults with:

Cervical or thoracic radiculopathy

Cauda equina syndrome

Neurogenic claudication or low back and leg symptoms related primarily to central spinal stenosis

Spondylolisthesis

Traumatic or congenital structural spinal abnormalities

Nonradicular leg or low back pain (i.e., discogenic or other nonspecific low back pain)

Intervention Surgical interventions for the treatment of radiculopathy, for example:

Laminectomy, laminotomy

Discectomy

Foraminotomy

Nucleotomy Includes “micro” approaches to the above procedures, which involve smaller incisions and/or areas of dissection and/or use of microscope or loupe magnification. Minimally invasive surgical procedures designed for treating radicular pain: percutaneous discectomy, discoplasty, nucleotomy, or nucleoplasty that are manual, automated, endoscopic, or laser-assisted, or use radiofrequency heat, including coblation technology.

Interventions involving combinations of the above interventions are eligible.

Surgical interventions primarily designed to treat neurogenic claudication and central spinal stenosis, spinal instability, or nonradicular low back pain, for example:

Spinal fusion

Arthroplasty

Artificial disc replacement

Interspinous process decompression (e.g., X-STOP® IPD System,8 Coflex® Interlaminar Technology)9

Minimally invasive lumbar decompression (mild® procedure)10

Other minimally invasive procedures designed for treating discogenic (i.e., nonradicular) low back pain

Epidural, spinal, or disc injections of enzymatic (e.g., chymopapain), chemical, or biologic (e.g., stem cells, mesenchymal cells) agents. Interventions involving combinations of procedures that include an above intervention are ineligible.

Comparator Placebo or no treatment comparators: sham surgery, expectant management, no treatment Active treatment comparators:

Nonsurgical management (e.g., physical therapy, exercise, pharmacologic treatment of symptoms, spinal manipulation, chiropractic treatment, epidural steroid or pain injections, other noninvasive treatments)

Surgical interventions as listed under “intervention”

No comparator

Chemonucleolysis Studies using “usual care” comparator groups will not be excluded but will be synthesized separately if no information was provided about the components of “usual care.”.

Outcomes Efficacy (at 4 weeks post-op or later):

Pain

Physical functioning

Social functioning

Psychological/emotional distress


Neurologic symptoms (e.g., weakness, sensory alteration)

Return to work

Reoperations for relapses/recurrent symptoms

Other outcomes not specifically listed as eligible. Pain, quality of life, and functional outcomes not measured using valid and reliable instruments or scales.11,12

(continued)



Table ES-2. Population, intervention, comparator, outcome, timing, setting and other study selection criteria for HTA on surgery for lumbar radiculopathy (continued)


Outcomes (continued)

Safety:

Surgery-related morbidity (e.g., venous thromboembolism, paralysis, new neurological symptoms, dural tear, epidural hematoma)

Surgical mortality (30 day)



Cost-effectiveness: Cost per quality-adjusted life year gained

Cost per disability-adjusted life year gained

Setting Inpatient or outpatient settings in countries categorized as “very high” on United Nations Human Development Index

Studies conducted in countries not categorized as “very high” on United Nations Human Development index.

Study Design and Risk of Bias Rating

For all Efficacy and Safety Research Questions: CCTs, RCTs, and SRs of CCTs or RCTs with similar scope as this HTA. For studies using active comparators, only RCTs or SRs of RCTs will be included. For Cost-Effectiveness Questions: CEA, CUA, or CBA performed from the societal or payer perspective For All Studies: Any risk of bias rating, but high risk of bias studies will only be used in quantitative syntheses if fewer than 3 studies are available.

Editorials, comments, letters, narrative reviews, case reports, case series, cohort studies, case-control studies.

Language and Time Period

English, no restrictions on time period included. Languages other than English.

Abbreviations: CCT = controlled clinical trial; CEA = cost-effectiveness analysis; CUA = cost-utility analysis; CBA = cost-

benefit analysis; HTA = health technology assessment; RCT = randomized controlled trial; SR = systematic review

ES-2.1.3 What is Excluded from This HTA

This review will not include studies published in languages other than English on conducted in

countries that are considered not very highly developed (based on UN Human Development

Programme).13 We also exclude studies evaluating surgical interventions performed primarily to

manage symptoms of central spinal canal stenosis (e.g., neurogenic claudication),

spondylolisthesis, traumatic or congenital abnormalities, or radiculopathy resulting from

infectious or neoplastic processes. This review will also not include studies conducted among

children or adolescents and will not cover surgical interventions for chronic discogenic low back

pain that is not radicular. Finally, this review excludes observational study designs (e.g., case

series, comparative cohort studies).

ES-2.1.4 Data Abstraction and Quality Assessment

One team member extracted relevant study data into a structured abstraction form and the lead

investigator checked it for accuracy. We used the Cochrane Risk of Bias (RoB 2.0) tool to assess

the risk of bias for each included study.14 Domains assessed with this tool include: bias arising

from randomization process, bias due to deviations from intended interventions, bias due to



missing outcome data, bias in measurement of the outcome, and bias in selection of the reported

result. Risk of bias was assessed as “high,” “some concerns,” or “low” at the study level unless

different outcomes within a single study required different risk of bias ratings. Two team

members conducted independent risk of bias assessments on all included studies.

ES-2.1.5 Data Synthesis and Analysis

Study characteristics and results were qualitatively synthesized for each research question in

tabular and narrative formats. We synthesized studies comparing the surgical interventions to

nonsurgical interventions separately from studies comparing alternative surgical interventions.

We summarized continuous outcome measures as absolute mean differences (AMD) wherever

possible. We summarized categorical outcomes using proportions. We transformed cost

outcomes in foreign currency to U.S. dollars based on the U.S. Department of Treasury mid-year

exchange rate for the year reported by study authors and then used the chain-weighted consumer

price index (CPI) to adjust to 2010 U.S. dollars.15,16 We required three or more studies with

similar intervention and comparator with same outcome measure at approximately the same

follow-up time point to calculate a pooled effect estimate. We considered outcomes reported at

less than 12 weeks to be short-term, outcomes reported between 12 weeks up to 52 weeks as

medium-term, and outcomes reported at 52 weeks or later as long-term. We estimated pooled

effects using a random effects model with the ‘metafor’ package in R using the DerSimonian and

Laird method.17

We graded the strength of evidence for each research question and each outcome measure using

GRADE, which assesses the strength of evidence based on domains relating to risk of bias,

inconsistency, imprecision, indirectness, and other considerations, such as publication bias.18

With GRADE, the strength of evidence can be graded as “very low,” “low,” “moderate,” or

“high”, and this rating represents the overall certainty of the findings. An evidence base

consisting of RCTs begins with a high strength of evidence rating based on study design but can

be downgraded to moderate, low, or very low based on serious or very serious concerns in the

domains assessed.

ES-2.2 Clinical Practice Guideline Synthesis

In addition to the systematic evidence review portion of this HTA, we also identified relevant

clinical practice guidelines and conducted a quality assessment of each guideline using the

Appraisal of Guidelines for Research & Evaluation II instrument.19,20 With this instrument, six

domains are assessed and an overall score of between 1 (lowest possible) and 7 (highest

possible) are assigned to reflect the overall quality of the guideline. We synthesized clinical

practice guidelines in a tabular format.

ES-3. Results

ES-3.1 Literature Yield

We identified and screened 1,860 unique citations. We excluded 1,638 citations after title and

abstract review. We reviewed the full-text of 222 articles and included a total of 25 studies

reported in 38 articles published between 1983 and 2017. Twenty-two RCTs provided evidence

on efficacy or comparative effectiveness (EQ1), two RCTs provided evidence on the



effectiveness or comparative effectiveness of revision surgical interventions for relapse (EQ2),

24 RCTs provided evidence on safety (SQ1), and seven studies (six RCTs and one cost-

effectiveness analysis) provided evidence on costs or cost-effectiveness (CQ1). The Full Report

describes individual study and population characteristics and findings for all included studies

(Appendix C), the list of studies we screened but excluded at the full-text stage (Appendix D),

and risk of bias assessments for included studies (Appendix E).

ES-3.2 Efficacy

ES-3.2.1 Efficacy Question 1

In adults with symptomatic lumbar radiculopathy, what is the effectiveness and comparative


We included 22 RCTs. Four were conducted in the U.S.;21 the rest were conducted in Canada

(N=1),22 Taiwan (N=1),23 Japan (N=1)24 or various European countries (N=15).5,25-38 Seven

RCTs provided evidence for the efficacy of surgery compared with nonsurgical

treatment;5,21,22,30,31,35,39 fifteen RCTs (providing 17 comparisons) provided evidence for the

comparative effectiveness of alternative surgical interventions.23-29,33,34,36-38,40-42 The

interventions and comparators evaluated in studies are summarized in Table ES-3. Across the

included RCTs, studies reported outcomes at various time points spanning from immediately

postoperative to up to 10 years postoperative; no single efficacy measure was used consistently

across all included studies. We rated one RCT as low risk of bias,43 10 RCTs as some concerns

for bias,22,23,25-27,33,34,36,37,41 and 10 RCTs as high risk of bias.5,21,24,28-32,35,40

Table ES-3. Surgical and comparator interventions used among 22 included studies for EQ1

Surgical Interventiona Comparator Interventionb

Eff

icac

y R

CT

s

(k=

7)

Microdiscectomy Spinal manipulation22; Physiotherapy31

Percutaneous disc decompression with coblation technology39

Epidural steroid injection

Percutaneous disc decompression35 Discectomy5,30 Discectomy/microdiscectomy21

Conservative management

Co

mp

arat

ive

effe

ctiv

enes

s

RC

Ts

(k=

15)

Tubular/trocar discectomy28,38 Automated percutaneous lumbar discectomy36 Percutaneous endoscopic discectomy32 Endoscopic interlaminar or transforaminal discectomy29 Microendoscopic discectomy24,27 Sequestrectomy26 Percutaneous laser disc decompression37 Microscopically assisted percutaneous nucleotomy34

Microdiscectomy

Automated percutaneous discectomy/endoscopic percutaneous discectomy40 Video-assisted arthroscopic microdiscectomy41 cMicroendoscopic discectomy23,27

Discectomy

cMicrodiscectomy25,27,33 Discectomy a In the Appendix C Evidence Tables, these interventions are considered the surgical group and are denoted as SG1.

b In the Appendix C Evidence Tables, these interventions are considered the comparator groups; surgical comparators are denoted

as SG2 or SG3. Nonsurgical comparator groups are denoted as NS1.

c This study was a three-arm RCT that allocated participants to microendoscopic discectomy, microdiscectomy, and standard

discectomy; thus, it contributes to three comparisons of interest for this HTA.



Abbreviations: RCT = randomized controlled trial.

A. Surgery compared with nonsurgical interventions

Seven RCTs that compared surgery to non-surgical interventions reported at least one efficacy

outcome.5,21,22,30,31,35,39 Five were rated as high risk of bias,5,21,30,31,35 one was rated as some

concerns for bias,22 and one was rated as high risk of bias for outcomes later than 12 weeks and

some concerns for bias for outcomes less than 12 weeks.39 The non-surgical interventions to

which surgery was compared included medications, physical therapy, patient

education/counseling, spinal manipulation, and epidural steroid injection. We were unable to

conduct quantitative synthesis for any outcomes because of outcome heterogeneity and because

some studies did not report measures of variance needed to conduct a meta-analysis. Table ES-4

summarizes findings and strength of evidence ratings. See the Full Report for complete details.

Table ES-4. Summary of efficacy outcome findings and strength of evidence ratings comparing surgery to nonsurgical interventions in persons with symptomatic lumbar radiculopathy (EQ1)

Outcomes Length of followup

No. studies (k) No. participants (N)

Summary of effect favors surgery; favors non-surgical intervention;

no difference Certaintya

Painb k=7; N=1,158

Up to 26 weeks k=5; N=970 Surgery reduces pain more than non-surgical interventions by an amount considered a minimally-important difference for most measures reported.

⨁⨁◯◯

LOW

Between 1 and 8 years k=3; N=840 Surgery and non-surgical interventions decrease pain by about the same amount.

⨁◯◯◯

VERY LOW

Function/Disabilityc k=5; N=970

Up to 26 weeks k=4; N=880 Surgery improves function and reduces disability more than non-surgical interventions by an amount considered a minimally-important difference for most measures reported.

⨁◯◯◯

VERY LOW

Between 1 and 8 years k=4; N=930 Surgery and non-surgical interventions improve function and reduce disability by about the same amount.

⨁◯◯◯

VERY LOW

Quality of lifed k=2; N=96

Up to 12 weeks k=1; N=40

Surgery and non-surgical interventions improve quality of life by about the same amount.

⨁◯◯◯

VERY LOW Between 12 weeks and 2 years

k=1; N=56

Neurologic symptomse 6 weeks to 52 weeks

k=2; N=146 Surgery and non-surgical interventions improve neurologic symptoms by about the same amount.

⨁◯◯◯

VERY LOW

Return to workf Between 12 weeks and 10 years

k=5; N=835 Return to work outcomes are similar for surgery and non-surgical interventions.

⨁◯◯◯

VERY LOW

a We assessed certainty using GRADE, which assesses the evidence base for each outcome measure based on risk of bias,

inconsistency, imprecision, indirectness, and other considerations; certainty is rated as “very low”, “low”, “moderate”, or “high”.

For domains with more than 1 measure reported (e.g., pain), we rated each measure separately but this table only depicts the most

common level associated with the domain. See Full Report for details.

b As measured by visual analogue scale (VAS) for leg pain and for back pain, SF-36 Bodily Pain subscale, Sciatica Index, and

other patient-reported measures of pain.

c As measured by Oswestry Disability Index, Roland-Morris Disability Questionnaire, and SF-36 Physical Functioning subscale.

d As measured by SF-36 and 15D health-related quality of life measures.

e As measured by physical exam or patient-report.

f As measured by actual return to work, self-reported ability to work, receipt of disability benefits, or other related measures.



Pain

All seven RCTs reported as least one pain outcome. Pain outcomes reported included the Visual

Analogue Scale (VAS) 100 mm or 10 cm for leg pain, the VAS 100 mm or 10 cm for back pain,

the SF-36 bodily pain subscale, the Sciatica index, the McGill Pain Questionnaire, the

Abderdeen back pain scale. A few studies also reported the frequency and proportion of

participants reporting reduced pain, no pain, or relief from pain. Peul et al.,30 Osterman et al.,31

and Gerszten et al.,39 reported decreased VAS scores (i.e., improvement) for leg pain in

participants allocated to both the surgical treatment and nonsurgical comparator from baseline

through short-term (6 and 8-week followup) and medium-term (up to 26 weeks) followup.

Between-group differences in VAS scores ranged from -6 to -26 points favoring surgery at short-

and medium-term followup. Of the 2 studies reporting long-term outcomes, improvements in

pain persisted but between-group differences were not significant.

Function and Disability

Five RCTs reported various measures of physical, mental, emotional, and social functioning or

disability.21,22,30,31,39 Functional outcomes reported include the Oswestry disability index, the

physical functioning subscale of the SF-36, the Roland-Morris Disability Questionnaire, the

Prolo Scale, and various other subscales of the SF-36. Across studies, function improved in both

participants allocated to surgery and in participants allocated to non-surgical interventions;

between-group differences favored surgery in the short-term, but differences did not persist in

the long-term. For example, Weinstein et al. [SPORT] reported the Oswestry Disability Index at

12 weeks, 52 weeks, 2 years, 4 years, and 8 years.21,44,45 Only the outcome reported at 12 weeks

demonstrated a significant difference between treatment groups, favoring surgery (AMD -4.7

[95% CI, -9.3 to -0.2]). Peul et al. reported outcomes using the Roland-Morris Disability

Questionnaire.30 At 8 weeks, a larger improvement was observed in participants allocated to

microdiscectomy compared with participants allocated to conservative management (AMD -3.1

[95% CI, -4.3 to -1.7]). This difference between groups did not persist at 26 weeks, 52 weeks, 2

years, or 5 years.

Quality of life

Two RCTs reported health-related QOL outcomes. These studies reported outcomes using the

total SF-36 score22 (sum of all normed subscales, possible range 0 to 800) and the 15D QOL

measure (range 0 to 1.0).31 In both studies, QOL improved from baseline to followup in the

surgery and nonsurgical intervention groups; no significant differences between groups were

observed by either study.


Two RCTs reported outcomes related to neurological symptoms, specifically sensory or motor

deficits.31,39 With few exceptions, no between-group differences were observed. For example,

Osterman et al. reported a similar proportion of participants with muscle weakness among those

allocated to microdiscectomy (N=28) compared with those allocated to physiotherapy (N=28),

respectively, at 6 weeks (53.8% vs. 46.2%), 12 weeks (42.3% vs. 46.2%), and 52 weeks (28.6%

vs. 30%).31



Return to work

Five RCTs reported various outcomes related to “return to work.”5,21,31,35,39 Some measures

captured actual return to work, whereas others reflected somewhat indirect measures, such as

self-reported ability to work, receipt of disability benefits, or pain affecting occupational status.

With one exception35; no between-group differences in return to work outcomes were observed.

For example, Weinstein et al.[SPORT]21 (N=501) reported the difference in proportion of

participants working full time at 2 years followup between the participants allocated to surgery

compared with the participants allocated to conservative management was -2.2% (95% CI, -

10.6% to 6.2%).

Other efficacy outcomes

Four RCTs reported other efficacy outcomes related to perceived recovery, overall time to

recovery, overall result, and patient satisfaction with symptoms.5,21,30,31 These outcomes were

consistent with previously reported efficacy outcomes that suggest more favorable outcomes for

participants who are allocated to surgery in the short and medium term. Results from three of

these studies also suggest some favorable outcomes in the long-term. We did not use these

outcomes in our strength of evidence ratings because of heterogeneity in outcome definition.

B. Minimally invasive surgery compared with microdiscectomy or discectomy

Ten RCTs comparing minimally invasive surgical interventions (tubular/trocar discectomy,28,38

percutaneous endoscopic discectomy,32 endoscopic interlaminar or transforaminal discectomy,29

microendoscopic discectomy,24,27 sequestrectomy,26 percutaneous laser disc decompression,46

microscopically assisted percutaneous nucleotomy,34 and video-assisted microdiscectomy41)

reported at least one efficacy outcome. Four were rated as high risk of bias,24,28,29,32 five were

rated as having some concerns for bias,26,27,34,37,41 and one was rated as low risk for bias.38 Table

ES-5 summarizes findings and strength of evidence ratings. See the Full Report for complete

details.

Pain

All 10 RCTs reported at least one pain outcome. Two studies reported decreases in VAS 100 mm

leg and back pain scores from baseline to 4 weeks and 8 weeks followup among participants

allocated to minimally invasive surgery and among participants allocated to

microdiscectomy.37,38 Between-group differences in short-term outcomes were not significant in

either study for either leg or back pain. For example, Arts et al. reported between group

differences of 4.5 (95% CI, -3 to 9.3) at 4 weeks and 4.5 (95% CI, -0.4 to 9.3) at 8 weeks for

VAS 100 mm leg pain scores.38

Five RCTs reported medium-term outcomes using VAS 100 mm or 10 cm for leg and back pain.

With one exception, between-group differences were not significant in any study. The pooled

between-group difference in VAS 100 mm leg pain at 12 to 26 weeks was 0.3 (95% CI, -2.2 to

2.9, 4 RCTs, 642 participants, I2=0%, Appendix F, Figure F-1). The pooled between-group

difference in VAS 100 mm back pain at 12 to 26 weeks was 1.3 (95% CI, -3.5 to 6.2, 4 RCTs,

642 participants, I2=61.7%, Appendix F, Figure F-2).



Table ES-5. Summary of efficacy outcome findings and strength of evidence ratings comparing minimally-invasive surgery to standard surgery in persons with symptomatic lumbar radiculopathy (EQ1)



Summary of effect favors minimally-invasive surgery; favors standard

surgery; no difference Certaintya

Painb k=10; N=1,155

Up to 26 weeks k=5; N=869 Improvements in pain are similar between minimally-invasive surgery and standard surgery.

⨁⨁⨁◯

MODERATE to LOW

52 weeks to 2 years k=5; N=869 Improvements in pain are similar between minimally-invasive surgery and standard surgery.

⨁⨁◯◯

LOW

Function/Disabilityc k=8; N=1,063

Up to 26 weeks k=6; N=903 Improvements in function are similar between minimally-invasive surgery and standard surgery.

⨁⨁◯◯

LOW to VERY LOW

52 weeks to 2 years k=8; N=1,063 Improvements in function are similar between minimally-invasive surgery and standard surgery.

⨁◯◯◯

VERY LOW

Quality of lifed

12 weeks to 3 years k=3; N=286

Quality of life improvements are similar between minimally-invasive surgery and standard surgery.

⨁◯◯◯

VERY LOW

Neurologic symptomse

12 weeks to 2 years k=6; N=602 Neurologic symptom improvements are similar between

minimally-invasive surgery and standard surgery. ⨁◯◯◯

VERY LOW

Return to workf K=6; N= 555 Minimally-invasive surgery reduces the duration of postoperative disability by a range of 4 weeks to 15 weeks compared to standard surgery.

⨁◯◯◯

VERY LOW





b Only studies that assessed pain as measured by visual analogue scale (VAS) for leg pain and for back pain, SF-36 Bodily Pain

subscale, and Sciatica Index were included in strength of evidence ratings.

c Only studies that assessed pain as measured by Oswestry Disability Index, Roland-Morris Disability Questionnaire, and SF-36

Physical Functioning subscale, and Prolo scale were included in strength of evidence ratings.

d As measured by the physical health and mental health component summary scores of the SF-36.

e As measured by physical exam or patient-report.

f As measured by mean duration of post-operative disability or self-reported “work impairment”.

These 5 RCTs also reported long-term outcomes at 52 weeks and 2 years. The pooled between-

group difference in VAS 100 mm leg pain at 52 weeks to 1.5 years was 1.6 (95% CI, -1.5 to 4.6,

4 RCTs, 640 participants, I2=28.1%) and the pooled between-group difference in VAS 100 mm

back pain at 52 weeks to 1.5 years was 1.5 (95% CI, -3.0 to 5.9, 4 RCTs, 640 participants,

I2=57.6%). Four RCTs reported pain outcomes using the bodily pain subscale of the SF-

36.26,28,37,38 In all studies, pain scores improved from baseline to short-, medium-, and long-term

followup among participants allocated to both surgical groups. Increases in scores ranged from

37 to 51 points over the various follow-up times. With one exception, no between-group

differences were observed at any follow-up time. The pooled mean difference in SF-36 bodily

pain scores at 12 to 26 weeks was -3.0 (95 % CI, -12.8 to 6.8, 3 RCTs, 500 participants,

I2=75.4%, Appendix F, Figure F-3).



Function and disability

Eight RCTs reported at least one outcome related to functioning or disability26-29,34,37,38,40

Functional outcomes reported included the Oswestry disability index, the Roland-Morris

Disability Questionnaire, the physical functioning subscale of the SF-36, the Prolo Scale, and

various other subscales of the SF-36. With few exceptions, between-group differences were

minimal. For example, Ryang et al. reported decreases of 41 points on the Oswestry Disability

Index among participants allocated to trocar microdiscectomy and 44.7 points among

participants allocated to microdiscectomy at 1.3 years (P=0.83, calculated AMD 3.6).28 Haines et

al. reported no significant between-group difference (P=0.74) in change from baseline scores of

the Roland-Morris Disability Questionnaire at 26 weeks (calculated AMD 0.02).40 Five RCTs

reported outcomes with the SF-36 physical functioning subscale.26,28,38,40 Between-group

differences favoring minimally invasive surgery at 4 weeks were observed by one study.37 This

difference did not persist at 8 weeks. In the medium-term (12 weeks to 26 weeks), the pooled

between-group mean difference in SF-36 physical functioning subscale was -2.4 (95 % CI, -6.1

to 1.2), 4 RCTs, 527 participants, I2=0.0%, Appendix F, Figure-F4). Some between-group

differences were observed at 52 weeks and 2 years; but the findings were mixed with respect to

which group was favored.

Quality of life

Three RCTs reported health-related QOL using (PCS) and mental health (MCS) component

summary scores over 12 weeks to 2.8 years.26-28 In all studies, quality of life as measured by both

component scores improved over time in both intervention groups, and with one exception, no

statistically significant between-group differences were observed. Ryang et al. observed a

significant difference in the SF-MCS at the 1.3 year followup; participants allocated to

microdiscectomy had a higher score (mean 51.9 [SD 7.8]) compared with participants allocated

to minimal access trocar microdiscectomy (mean 44.0 [SD 13.2], P=0.03), but it is not clear

whether this comparison adjusted for small differences in baseline scores.28


Six RCTs comparing minimally invasive surgery to microdiscectomy26,28,29,32,34 and discectomy41

reported outcomes related to neurological symptoms. Findings were not reported by group in one

RCT,34 the remaining five studies observed no between-group differences. Three RCTs reported

no statistical difference in neurological symptoms between intervention groups.26,28,29 We

calculated no differences in the other 2 RCTs.32,41 For example, Ryang et al. reported no

difference in the proportion of participants with sensory deficits (40% vs. 43%) in participants

allocated to minimal access trocar discectomy compared with participants allocated to

microdiscectomy, respectively, over an average of 1.3 years followup (P=0.31).28 Similar

findings were observed for the proportion with motor deficits (27% vs. 23%, P=0.86).

Return to work

Six RCTs reported various outcomes related to “return to work”, though in some studies this

outcome was not reported by group.26,29,32,34,36,41 Of the 4 RCTs that reported between-group

differences, three RCTs29,32,41 suggest that participants allocated to minimally-invasive surgery

return to work sooner than participants allocated to standard surgery as measured by weeks of

postoperative disability. The range of this difference was 4 weeks to 15 weeks. The remaining

RCT26 reported no significant between-group differences; however; this study used a multi-level



categorical measure of work impairment, which may be measuring a related, but different

construct compared to the other three RCTs. In this study, Thome et al. reported specific

categories of impairment of work at 12 to 26 weeks and at 2 years.26 Thirty-one percent of

participants allocated to sequestrectomy reported that their work impairment was “much better”

at 12 to 26 weeks compared with 33% of participants allocated to microdiscectomy. At 2 years,

the proportions were 37% and 31%, respectively. The proportion of participants endorsing

various categories of work impairment were not significantly different between groups (P=0.415

at 12 to 26 weeks, P=0.112 at 2 years).


Ten RCTs reported other efficacy outcomes, related to perceived recovery, overall time to

recovery, overall result, and patient satisfaction with symptoms. With few exceptions, most

observed no significant differences between groups. We did not use these outcomes in our

strength of evidence ratings because of heterogeneity in outcome definition.

C. Microdiscectomy compared to discectomy

Three RCTs comparing microdiscectomy to discectomy reported efficacy outcomes.25,27,33 All

three were rated as some concerns for bias. Table ES-6 summarizes findings and strength of

evidence ratings. See the Full Report for complete details.

Table ES-6. Summary of efficacy outcome findings and strength of evidence ratings comparing microdiscectomy to discectomy in persons with symptomatic lumbar radiculopathy (EQ1)



Summary of effect favors microdiscectomy; favors discectomy;


Painb k=3; N=282

Up to 6 weeks k=1; N=80 Pain decreased by similar amounts in both surgical groups.

⨁◯◯◯

VERY LOW

26 weeks to 2 years k=2; N=202 Pain decreased by similar amounts in both surgical groups.

⨁◯◯◯

VERY LOW

Function/Disabilityc


Function improved by similar amounts in both surgical groups.

⨁⨁◯◯

LOW

Quality of lifed


Health-related quality of life improved by similar amounts in both surgical groups.

⨁⨁◯◯

LOW

Neurologic symptomse k=0

Return to worke k=1; N=60 Both surgeries result in similar duration of postoperative work disability; 10.4 weeks for microdiscectomy compared with 10.1 weeks for discectomy.

⨁◯◯◯

VERY LOW





b As measured by visual analogue scale (VAS) for leg pain and for back pain.

c As measured by Oswestry Disability Index.

d As measured by the physical health and mental health component summary scores of the SF-36.

e As measured by duration of postoperative disability and by the proportion out of work at unspecified followup time.



Pain

Three trials reported pain outcomes using a VAS 10 cm scale at 4 and 6 weeks,33 at 52 weeks,25

and at 26 weeks, 52 weeks, and 2 years.27 In Henriksen et al., actual VAS values were not

reported but no differences were reported between groups for both VAS leg pain and VAS back

pain at 4 weeks and at 6 weeks followup.33 Tullberg et al. reported a mean baseline VAS 10 cm

leg pain score of 7.0 (SD NR) among participants allocated to microdiscectomy and 7.0 (SD NR)

among participants allocated to discectomy.25 The mean scores at 52 weeks were 2.1 (SD NR)

and 2.3 (SD NR), respectively (AMDs and P value NR). Teli et al. reported VAS 10 cm leg pain

score of 8 (SD 1) at baseline decreasing to 2 (1 SD) at 26 weeks, 1(1) at 52 weeks, and 2 (1) at 2

years in both surgical groups (P=0.73 for between-group differences).27

Function and disability

One RCT reported outcomes with the Oswestry Disability Index.27 Teli et al. observed scores

improve at 26 weeks, 52 weeks, and 2 years among participants allocated to microdiscectomy

and in participants allocated to discectomy. Among participants allocated to microdiscectomy,

score decreases from baseline (40 [SD 4]) ranged from 25 to 29 points; among participants

allocated to discectomy score decreases from baseline (39 [SD 4]) ranged from 24 to 27 points.

No significant between-group differences were observed (P=0.81).

Quality of life

One RCT reported outcomes with the SF-36 physical health component summary (PCS) score

and the mental health component summary (MCS) score at 26 weeks, 52 weeks, and 2 years.27

For PCS, Teli et al. reported increases from baseline (21 [SD 4]) ranging from 19 to 23 points

among participants allocated to microdiscectomy at the various follow-up time points compared

with increases from baseline (22 [SD 4]) ranging from 18 to 22 points among participants

allocated to discectomy. No significant between-group differences were observed (P=0.68).

Similar findings were reported for the MCS (P=0.78 for between-group differences).


No studies reported outcomes related to neurological symptoms.

Return to work

One RCT reported on outcomes related to “return to work”.25 Tullberg et al. reported a mean

duration of postoperative, full-time sick leave of 10.4 (SD NR) weeks in participants allocated to

microdiscectomy compared with 10.1 (SD NR) weeks in participants allocated to discectomy (P

value NR). The proportion out of work at an unspecified follow-up time point was 16.7% among

those allocated to microdiscectomy and 6.7% among those allocated to discectomy (calculated P

=0.42).


One RCT rated reported the frequency and proportion of participants with a specified opinion on

recovery at 52 weeks (total recovery, almost recovered, good, unchanged, or worse).25 Among

participants allocated to microdiscectomy, 11 (37.9%) reported total recovery, and 8 (27.6%)

reported almost recovered. Among participants allocated to discectomy these outcomes were 6

(20.7%) and 14 (28.3%), respectively (calculated P=0.25 and 0.18, respectively).



ES-3.2.2 Efficacy Question 2

In adults with symptomatic lumbar radiculopathy, does effectiveness or comparative

effectiveness of surgical interventions vary for patients who are not employed because of

disability or patients who are undergoing recurrent surgery for relapse?

We did not identify any studies that reported outcomes specifically for patients not employed

because of disability.

We identified two studies focused on the efficacy47 or comparative effectiveness48 of revision

surgery for relapse; both were rated as high risk of bias. North et al. was conducted in the United

States and randomized 50 participants with persistent radicular pain despite one or more prior

lumbosacral spine surgeries to either repeat lumbosacral decompression or spinal cord

stimulation.47 Ruetten et al. was conducted in Germany and randomized 100 adults who had a

previous conventional discectomy with acute occurrence of radicular leg symptoms after a pain-

free interval in combination with a recurrent disc herniation on MRI to either revision

endoscopic discectomy or revision microdiscectomy.48 Table ES-7 and Table ES-8 summarizes

findings and strength of evidence ratings for EQ2. See the Full Report for complete details.

Table ES-7. Summary of efficacy findings and strength of evidence ratings comparing repeat lumbosacral decompression surgery with spinal cord stimulation for treatment of lumbar radiculopathy relapses (EQ2)



Summary of effect favors repeat surgery; favors spinal cord stimulation;


Pain k=0

Function/disabilityb

1.8 to 5.7 years k=1; N=50

Similar levels of improvement in function and disability from repeat surgery and spinal cord stimulation.

⨁◯◯◯

VERY LOW

Quality of life k=0

Neurologic symptoms k=0

Return to workc 1.8 to 5.7 years

k=1; N=50 Return to work outcomes were similar between participants receiving repeat lumbosacral decompression and participants receiving spinal cord stimulation.

⨁◯◯◯

VERY LOW





b As measured by patient-reported impairment from pain in performing everyday activities.

c Specific measure used was poorly defined.



Table ES-8. Summary of efficacy findings and strength of evidence ratings comparing revision endoscopic discectomy with revision microdiscectomy for treatment of relapsed lumbar radiculopathy (EQ2)



Summary of effect favors revision endoscopic surgery; favors

microdiscectomy; no difference Certaintya

Painb 12 weeks to 2 years

k=1; N=100 Similar improvements in pain over time in both surgical groups.

⨁◯◯◯

VERY LOW

Function/disabilityc


Similar improvements in function and disability over time in both surgical groups.

⨁◯◯◯

VERY LOW

Quality of life K=0

Neurologic symptomsd 12 weeks to 2 years

k=1; N=100 Similar improvement in neurologic symptoms over time in both surgical groups.

⨁◯◯◯

VERY LOW

Return to worke k=1; N=100 Revision endoscopic surgery results in a shorter duration of postoperative disability (4 weeks) compared to microdiscectomy (7.4 weeks).

⨁◯◯◯

VERY LOW





b As measured by visual analogue scale 100 mm for leg and back pain, and North American Spine Society Pain score.

c As measured by Oswestry Disability Index.

d As measured by North American Spine Society Neurology score.

e As measured by mean duration of post-operative disability, P< 0.01.

Pain

North et al. did not report any outcomes related to pain. Ruetten et al. reported improvement in

VAS 100 mm leg pain score from baseline to 12 weeks, 26 weeks, 52 weeks, and 2 years among

participants allocated to revision endoscopic discectomy and among participants allocated to

revision microdiscectomy.48 Between-group differences were reported as not significant at any

follow-up time point (AMDs were NR). A similar pattern was observed for VAS 100 mm back

pain scores and North American Spine Society pain scores.

Functioning/disability

North et al. reported no significant differences in qualitative assessment of impairment.47 Ruetten

et al. reported improvements as measured by the Oswestry disability index.48 from baseline to 12

weeks, 26 weeks, 52 weeks, and 2 years among participants allocated to revision endoscopic

discectomy and among participants allocated to revision microdiscectomy. The between-group

differences were reported as not significant.

Quality of life

Neither study reported outcomes related to overall quality of life.


North et al. did not report any outcomes related to neurologic symptoms. Ruetten et al. reported

mean North American Spine Society Neurology scores at 12 weeks, 26 weeks, 52 weeks, and 2

years.48 Scores in both groups improved over time and the differences between groups were

reported as not significant.



Return to work

North et al. reported no significant differences in return to work, but actual values were not

reported. Ruetten et al. reported a significant difference between groups in the mean duration of

postoperative disability.48 Among participants allocated to revision endoscopic discectomy the

mean was 4 weeks (SD NR) and among participants allocated to revision microdiscectomy the

mean was 7.4 weeks (SD NR) (P < 0.01).


North et al. reported on the frequency and proportion of successful treatment over a mean

follow-up time of 2.9 years (range 1.8 years to 5.7 years).47 Success was defined as at least 50%

pain relief and patient satisfaction with treatment. A significant difference in treatment success

was observed (P < 0.01). Among those allocated to repeat lumbosacral decompression,

successful treatment was observed in 3 (12%). Among those allocated to spinal cord stimulation,

successful treatment was observed in 9 (47%). Ruetten et al. reported on the frequency and

proportion of patient satisfaction with surgery and whether participants would undergo the

operation again. Among those allocated to revision endoscopic discectomy, 43 (95%) were

satisfied; among those allocated to revision microdiscectomy 36 (86%) were satisfied (P value

for comparison NR).

ES-3.3 Safety

Safety Question 1

In adults with symptomatic lumbar radiculopathy, what are the adverse events associated with

surgical interventions?

All 24 RCTs included for EQ1 and the two RCTs included for EQ2 also provided evidence for

safety outcomes. The study characteristics for these RCTs have been previously described.


Seven RCTs that compared surgery to non-surgical interventions reported at least one safety

outcome.5,21,22,30,31,35,39 Table ES-9 summarizes findings and strength of evidence ratings. See the

Full Report for complete details.



Table ES-9. Summary of safety outcome findings and strength of evidence ratings comparing surgery to nonsurgical interventions in persons with symptomatic lumbar radiculopathy (SQ1)



Summary of effect favors surgery; favors non-surgical intervention;


Surgical mortality k=6; N=1,096 NAb Surgical mortality is rare; no deaths reported among participants allocated to surgery in any studies.

⨁⨁◯◯

LOW

All-cause mortality Up to 10 years

k=3; N=717 All-cause mortality is rare and is similar for surgery and non-surgical interventions.

⨁⨁◯◯

LOW

Surgical morbidity k=6; N=1,032 NAb Surgical morbidity occurs with low frequency; dural tears are the most common adverse event (reported in up to 4% of cases).

⨁⨁◯◯

LOW

Reoperations Up to 5 years

k=5; N=942 NAb The incidence of reoperations varies from 0% to 10.1%. ⨁◯◯◯

VERY LOW

Persistent opioid use Up to 26 weeks

K=1; N=90 Surgery and non-surgical interventions result in similar frequency of persistent opioid use.

⨁◯◯◯

VERY LOW a We assessed certainty using GRADE, which assesses the evidence base for each outcome measure based on risk of bias,


b Not applicable for comparative assessment because comparator treatment is non-surgical intervention.

Mortality

Six RCTS reported on mortality.5,21,22,30,31,39 All of these studies were rated as low risk of bias for

this specific outcome. Surgical mortality is not relevant as a comparative outcome given the non-

surgical comparison group. Thus, the strength of evidence for surgical mortality reflects our

certainty about the absolute incidence of surgical mortality in the surgical intervention group. Of

the RCTs that reported surgical mortality, no studies reported any deaths relating to percutaneous

disc decompression,39 microdiscectomy,22,31 discectomy,30 or discectomy/microdiscectomy

procedures.21 Three RCTs reported all-cause mortality.5,21,39 For example, Weber et al. reported

three deaths (5.0%) among participants allocated to discectomy and no deaths among

participants allocated to conservative management at 10 years.5 One death was due to cancer and

two due to heart disease.

Surgical Morbidity

Six RCTs reported surgical morbidity outcomes.21,22,30,31,35,39 All of these studies were rated as

low risk of bias for this specific outcome. Surgical morbidity is not relevant as a comparative

outcome given the non-surgical comparison group. Thus, the strength of evidence for surgical

morbidity reflects our certainty about the absolute incidence of surgical morbidity in the surgical

intervention group. Surgical complications were generally rare among participants who

underwent surgical intervention. In the largest of the trials, Weinstein et al. [SPORT] reported 10

(4.0%) dural tear or spinal fluid leaks, 4 (1.6%) superficial postoperative wound infection, 1

(0.40%) vascular injury, 2 (0.81%) other intraoperative complications, and 9 (3.6%) other

unspecified postoperative complications among participants who underwent microdiscectomy.21

Reoperations

Five RCTs reported reoperation rates in participants that were allocated to and underwent the

surgical intervention; some studies also reported reoperations among participants who crossed

over from the nonsurgical intervention to surgery.21,22,30,31,35 Reoperations is not relevant as a

comparative outcome given the non-surgical comparison group. Thus, the strength of evidence



for reoperation reflects our certainty about the absolute incidence of reoperations among those

who underwent surgery, whether initially allocated to the surgical group or among those who

crossed over to surgery at some point during the trial. The incidence of reoperations across the

five RCTs varied from 0% to 10.1%. For example, Peul et al. reported that 7 (6%) participants

allocated to microdiscectomy had reoperations for recurrent sciatic within 2 years and 9 (7%) by

5 years.30 Among participants allocated to conservative management who crossed over to receive

surgery, 4 (6%) underwent a reoperation by 2 years and 8 (12%) by 5 years.

Persistent Opioid use

Only one RCT reported outcomes related to persistent opioid use.39 Gerszten et al. reported that

reduction in use of narcotics was not significantly different between participants who underwent

percutaneous disc decompression and those who underwent conservative management

participants at 26 weeks (actual values NR, P value NR).


Twelve RCTs that compared minimally-invasive surgery to non-surgical interventions reported

at least one safety outcome.23,26-29,32-34,36-38,41Table ES-10 summarizes findings and strength of

evidence ratings. See the Full Report for complete details.

Table ES-10. Summary of safety outcome findings and strength of evidence ratings comparing minimally-invasive surgery to standard surgery in persons with symptomatic lumbar radiculopathy (SQ1)



Summary of effect favors minimally-invasive surgery; favors standard

surgery; no difference

Certaintya

Surgical mortality k=5; N=464 No surgery-related deaths reported in any studies. ⨁⨁◯◯

LOW

All-cause mortality k=1; N=200 Only 1 death unrelated to surgery was reported. ⨁⨁◯◯

LOW

Surgical morbidity k=10; N=1,171

The most commonly reported complications were dural tears and spinal fluid leaks. Between-group differences were generally similar between groups with one exception.

⨁◯◯◯

VERY LOW

Reoperations k=10; N=1,058

The proportion of participants that had reoperations varied extensively across study groups (from 2.5 % to 64.5). Two studies observed significantly higher frequency of reoperations among participants who underwent minimally-invasive surgery compared to standard surgery.36,37 37,38 37,38 37,38 37,38 37,38 37,38

37,38 These findings were inconsistent with findings from the other 8 studies that showed a similar incidence.

⨁◯◯◯

VERY LOW

Persistent opioid use k=1; N=60 The duration of postoperative narcotic use ranged from 0.43 to 2 weeks (average 1 week) for participants who underwent video-assisted arthroscopic microdiscectomy and 1 to 8 weeks for participants who underwent discectomy.

⨁◯◯◯

VERY LOW



Mortality

Five RCTs reported mortality outcomes.23,27,29,32,41 No surgery-related deaths were reported. One

RCT reported all-cause mortality.29 Ruetten et al. reported one death (0.5%) unrelated to surgery;



the authors did not specify whether this death occurred among participants allocated to the

minimally invasive surgery or among participants allocated to microdiscectomy.

Surgical Morbidity

Ten RCTs23,26-29,32,34,37,38,41 reported surgical morbidity outcomes. The most common

complications reported were those relating to dural tear and spinal fluid leak. In nine of the 10

RCTs, morbidity incidence was similar between groups, though few reported statistical

significance testing. One RCT reported significantly fewer complications among participants

who underwent endoscopic discectomy compared to participants who underwent

microdiscectomy.29 In this study, Ruetten et al.29 reported significantly fewer complications

(P<0.05) among participants who underwent endoscopic discectomy compared to participants

who underwent microdiscectomy participants. Complication included transient postoperative

dysesthesia (3.3% vs 5.7%), postoperative bleeding (0% vs 2.3%), delayed wound healing (0%

vs 1.1%), and soft tissue infection (0% vs 1.1%).

Reoperations

Ten RCTs reported reoperation rates.26-29,32,34,36-38,41 The proportion of participants that had

reoperations varied extensively across study groups (from 2.5 % to 64.5). Two studies observed

significantly higher frequency of reoperations among participants who underwent minimally-

invasive surgery compared to standard surgery.36,37 These findings were inconsistent with

findings from the other 8 RCTs, which observed similar incidence of reoperations between

surgical groups.


One RCT reported outcomes related to persistent opioid use.41 Hermantin et al.41 reported the

duration of postoperative narcotic use ranged from 0.43 to 2 weeks (average 1 week) for

participants who underwent video-assisted arthroscopic microdiscectomy and 1 to 8 weeks

(average 3.65 weeks) for participants who underwent discectomy (P value NR).




Three RCTs that compared microdiscectomy to discectomy reported at least one safety

outcome.25,27,33 Table ES-11 summarizes findings and strength of evidence ratings. See the Full

Report for complete details.

Table ES-11. Summary of safety outcome findings and strength of evidence ratings comparing microdiscectomy to discectomy in persons with symptomatic lumbar radiculopathy (SQ1)



Summary of effect favors microdiscectomy; favors discectomy;


Surgical mortality k=1; N=142 No surgery-related deaths reported in either surgical group.

⨁⨁◯◯

LOW

All-cause mortality k=0

Surgical morbidity k=3; N=282 Surgical morbidity was infrequent and similar in both surgical groups.

⨁◯◯◯

VERY LOW

Reoperations

k=2; N= Incidence of reoperation similar in both surgical groups; range of incidence from 3% to 4.2%.

⨁◯◯◯

VERY LOW

Persistent opioid use k=0 a We assessed certainty using GRADE, which assesses the evidence base for each outcome measure based on risk of bias,


Mortality

Only one RCT reported a surgical mortality.27 Teli et al. reported no surgical deaths in either

group.27

Surgical Morbidity

Three RCTs reported surgical morbidity,25,27 but one33 did not report by group. In one RCT, the

overall frequency of surgical infection was 6.3%.41 The other two RCTs reported similar

frequency of complications between groups with respect to dural tears and nerve root injury.

Reoperations

Two RCTs reported on reoperations.25,27Tullberg et al. reported 1 (3.3%) reoperation by 52

weeks in each surgical group (microdiscectomy and discectomy).25 Teli et al. reported 3 (4.2%)

reoperations among participants who underwent microdiscectomy compared with 2 (3%) among

participants who underwent discectomy (calculated P=1.0).27


No RCTs reported persistent opioid use outcomes.



D. Revision surgery

Two RCTs that compared revision surgery to an alternative treatment reported at least one safety

outcome.47,48 Table ES-12 summarizes findings and strength of evidence ratings. See the Full


Table ES-12. Summary of safety outcome findings and strength of evidence ratings comparing revision surgery to spinal cord stimulation or microdiscectomy in persons with relapsed lumbar radiculopathy (SQ1)



Summary of effect favors revision surgerya; favors comparatora;

no difference Certaintyb

Surgical mortality k=2; N=150 No surgery-related deaths reported in any studies. ⨁⨁◯◯

LOW

All-cause mortality k=2; N=150 Only 1 death unrelated to surgery was reported. ⨁⨁◯◯

LOW

Surgical morbidity k=1; N=50 Surgical morbidity is infrequent; one site infection reported in participants who underwent spinal cord stimulation.

⨁◯◯◯

VERY LOW

k=1; N=100 Revision microendoscopic discectomy has significantly fewer serious complications compared to revision microdiscectomy group (6% vs 21%, P<0.05).

⨁◯◯◯

VERY LOW

Reoperations k=2; N=150 The incidence of reoperations ranged from 4% to 12.5%; no differences between groups

⨁◯◯◯

VERY LOW

Persistent opioid use At 2.9 years

k=1; N=50 A lower proportion of participants (58%) who underwent repeat surgery had stable or decreased opioid use compared with participants who underwent spinal cord stimulation at long-term followup (P=0.025).

⨁⨁◯◯

LOW

Unclear followup k-1; N=100 Postoperative opioid use was significantly reduced among participants who underwent microendoscopic discectomy compared with microdiscectomy.

⨁⨁◯◯

LOW

a For 1 RCT (North et al.47, the surgical intervention is repeat lumbosacral decompression and the comparator is spinal cord

stimulation. For 1 RCT (Ruetten et al.48, the surgical intervention is revision microendoscopic discectomy and the comparator is

standard microdiscectomy.

b We assessed certainty using GRADE, which assesses the evidence base for each outcome measure based on risk of bias,


ES 3.4 Cost

Cost Question 1

In adults with symptomatic lumbar radiculopathy, what is the cost-effectiveness of surgical

interventions?

We identified seven eligible studies reporting cost.27,42,49-53 Teli et al., a trial we also included for

efficacy and safety outcomes, reported on surgical costs of three alternative surgical

interventions.27 Five studies reported cost-effectiveness analyses related to RCTs that we also

included for efficacy and safety outcomes.49-53 Lastly, Malter et al. reported a cost-effectiveness

analysis using cost and effectiveness inputs from a variety of sources.42 Two studies were

conducted in the United States;42,50 the rest were conducted in the Netherlands,49,51,52 Italy,27 and

the United Kingdom.53 The time horizon used in studies ranged from 26 weeks to 10 years.



These studies reported cost findings using different currency and base years; thus, we converted

all figures to 2010 U.S. dollars (see Appendix B, Table B-1 for details on conversion) for this

report.


Three studies provided evidence for the cost-effectiveness of surgery compared with nonsurgical

treatment.42,49,50 Table ES-13 summarizes findings and strength of evidence ratings. See the Full


Table ES-13. Summary of cost-effective findings comparing surgery to nonsurgical interventions in persons with symptomatic lumbar radiculopathy (CQ1)



Summary of effect Certaintya

Cost-effectiveness Between 1 and 10 years

k=3; N=1,474b

Surgery results in higher quality-adjusted life years but similar or higher costs compared to non-surgical interventions. The mean cost per quality-adjusted life year gained from the payor perspective ranged from $51,156 to $83,322 (in 2010 dollars).

⨁◯◯◯

VERY LOW





b One study was a decision analysis not concurrent to an RCT so no N reported; and one study combined data from an RCT and a

concurrent observational study to contribute additional participants that were not included for efficacy or safety outcomes.


Four studies27,51-53 provided evidence for the comparative cost-effectiveness of alternative

surgical interventions, including percutaneous laser discectomy,52 tubular discectomy,51 and

automated percutaneous discectomy53 compared to microdiscectomy and a three-arm study

comparing microendoscopic discectomy, microdiscectomy, and discectomy.27 Table ES-14

summarizes findings and strength of evidence ratings. See the Full Report for complete details.

Table ES-14. Summary of cost-effective findings comparing minimally-invasive surgery to standard surgery in persons with symptomatic lumbar radiculopathy (CQ1)




Cost-effectiveness k=4; N=656

Inconsistent findings across studies. One study found higher surgical costs ($722) for minimally-invasive surgery compared to standard surgery; one study calculated a higher cost ($3573) per successful outcome for minimally-invasive surgery compared to standard surgery; one study reported no significant differences in quality-adjusted life years, total costs, or health care costs between groups, but point estimates suggest minimally-invasive surgery is less effective and costs more; and the last study reported no significant differences in quality-adjusted life years or total costs but some differences in health care costs suggesting minimally-invasive surgery costs less (AMD -$2,393 [95% CI, $-4,376 to $-409]).

⨁◯◯◯

VERY LOW






One study reported the surgical costs of microdiscectomy to discectomy, but did not report cost-

effectiveness.27 The cost of microdiscectomy was $3,156 (SD $438) and the cost of discectomy

was $2,976 (SD $322). Table ES-15 summarizes findings and strength of evidence ratings. See

the Full Report for complete details.

Table ES-15. Summary of costs comparing microdiscectomy to discectomy in persons with symptomatic lumbar radiculopathy (CQ1)




Costs k=1; N=142 Costs are slightly higher for microdiscectomy compared to discectomy ($3,156 (SD $438) vs $2,976 (SD $322). Calculated AMD $65 (95% CI, $52 to $307).

⨁⨁⨁◯

MODERATE



ES-3.5 Synthesis of Clinical Practice Guidelines

A synopsis of clinical practice guidelines (CPGs) and guidance related to the use of surgical

procedures for lumbar radiculopathy is summarized in Table ES-16. Please refer to the Full

Report for additional details. We assessed the quality of each CPG or procedure guidance using

the Appraisal of Guidelines for Research & Evaluation II instrument.19,20 With this instrument,

six domains are assessed and an overall score of between 1 (lowest possible) and 7 (highest

possible) are assigned to reflect the overall quality of the guideline.

Overall, the guidelines we identified were in general agreement about considering discectomy or

microdiscectomy (and related decompressive procedures) as acceptable treatment based on

evidence that it improves outcomes in the short to medium term. One guideline specifies that this

surgery can be considered when symptoms have not improved with conservative therapy.46

Another guideline suggests that conservative therapy is reasonable for patients with

nonprogressive symptoms who wish to delay surgery.54 The guideline recommendations relating

to minimally invasive spine surgery varied; one did not consider these specific procedures within

their scope.46 Three of the guidelines were developed 5 or more years ago; thus may not include

the most recent evidence for these procedures.55-57

Table ES-16. Synopsis of clinical practice guidelines related to lumbar radiculopathy or herniated intervertebral lumbar disc

Organization Guideline Title (Year) Guideline Qualitya

Synopsis of Recommendationb

American Pain Society Interventional Therapies, Surgery, and Interdisciplinary Rehabilitation for Low Back Pain (2009)55 Quality Rating: 5 out of 7

Open discectomy or microdiscectomy recommended for radiculopathy with prolapsed disc.

Insufficient evidence for determining superiority of open vs. micro approaches, and to evaluate alternative surgical methods including laser- or endoscopic-assisted techniques.

(continued)



Table ES-16. Synopsis of clinical practice guidelines related to lumbar radiculopathy or herniated intervertebral lumbar disc (continued)



National Institute for Health and Care Excellence (United Kingdom) Low back pain and sciatica in over 16s: assessment and management-Invasive treatments (2016)46 Quality Rating: 6 out of 7

Consider spinal decompression for sciatica (includes laminectomy, foraminotomy, and/or discectomy) when nonsurgical treatment has not improved pain or function and their radiological findings are consistent with sciatica symptoms.

American Society of Interventional Pain Physicians An Update of Comprehensive Evidence-Based Guidelines for Interventional Techniques in Chronic Spinal Pain (2013)20,56 Quality Rating: 4 out of 7

For lumbar disc prolapse, protrusion, and extrusion: automated percutaneous lumbar decompression (APLD), percutaneous lumbar disc decompression (PLDD), and mechanical decompression with nucleoplasty are recommended in select cases.

North American Spine Society Clinical Guidelines for Diagnosis and Treatment of Lumbar Disc Herniation with Radiculopathy (2012)57 Quality Rating: 5 out of 7

Discectomy is suggested to provide more effective symptom relief than medical/interventional care for patients with lumbar disc herniation with radiculopathy whose symptoms warrant surgical intervention. In patients with less severe symptoms, surgery or medical/interventional care appear to be effective for both short- and long-term relief.

Surgical intervention prior to 6 months is suggested in patients with symptomatic lumbar disc herniation whose symptoms are severe enough to warrant surgery. Earlier surgery (within 6 months to 1 year) is associated with faster recovery and improved long-term outcomes.

Use of an operative microscope is suggested to obtain comparable outcomes to open discectomy for patients with lumbar disc herniation with radiculopathy.

Endoscopic percutaneous discectomy is suggested for carefully selected patients to reduce early postoperative disability and reduce opioid use compared with open discectomy in the treatment of patients with lumbar disc herniation with radiculopathy.

In a select group of patients automated percutaneous lumbar discectomy (APLD) may achieve equivalent results to open discectomy, however, this equivalence is not felt to be generalizable to all patients with lumbar disc herniation with radiculopathy whose symptoms warrant surgery.

Insufficient evidence for other procedures (See Full Report for details)

American College of Occupational and Environmental Medicine Low Back Disorders. In Occupational medicine practice guidelines: evaluation and management of common health problems and functional recovery in workers (2016)54 Quality Rating: Unknownc

Patients with evidence of specific nerve root compromise confirmed by appropriate imaging studies may be expected to potentially benefit from surgery.

National Institute for Health and Care Excellence (United Kingdom) Percutaneous transforaminal endoscopic lumbar discectomy for sciatic: Interventional procedures guidance [IPG 556] (2016)]58 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous transforaminal endoscopic lumbar discectomy for sciatica is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit.

(continued)



Table ES-16. Synopsis of clinical practice guidelines related to lumbar radiculopathy or herniated intervertebral lumbar disc (continued)



National Institute for Health and Care Excellence (United Kingdom) Percutaneous interlaminar endoscopic lumbar discectomy for sciatica: Interventional procedures guidance[IPG555](2016)59 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous interlaminar endoscopic lumbar discectomy for sciatica is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit.

National Institute for Health and Care Excellence (United Kingdom) Percutaneous coblation of the intervertebral disc for low back pain and sciatica Interventional procedures guidance[IPG543](2016)60 Quality Rating: 2 out of 7

Current evidence on percutaneous coblation of the intervertebral disc for low back pain and sciatica raises no major safety concerns. The evidence on efficacy is adequate and includes large numbers of patients with appropriate follow-up periods. Therefore, this procedure may be used provided that normal arrangements are in place for clinical governance, consent and audit.

National Institute for Health and Care Excellence (United Kingdom) Percutaneous electrothermal treatment of the intervertebral disc annulus for low back pain and sciatica Interventional procedures guidance[IPG544](2016)61 Quality Rating: 2 out of 7

Current evidence on percutaneous electrothermal treatment of the intervertebral disc annulus for low back pain and sciatica raises no major safety concerns. The evidence on efficacy is inconsistent and of poor quality. Therefore, this procedure should only be used with special arrangements for clinical governance, consent and audit or research.

National Institute for Health and Care Excellence (United Kingdom) Percutaneous intradiscal radiofrequency treatment of the intervertebral disc nucleus for low back pain. Interventional procedures guidance[IPG545] (2016)62 Quality Rating: 2 out of 7

Current evidence on percutaneous intradiscal radiofrequency treatment of the intervertebral disc nucleus for low back pain raises no major safety concerns. The evidence on its efficacy is limited in quantity and quality. Therefore, this procedure should only be used with special arrangements for clinical governance, consent and audit or research.

National Institute for Health and Care Excellence (United Kingdom) Epiduroscopic lumbar discectomy through the sacral hiatus for sciatica Interventional procedures guidance[IPG570] (2016)63 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of epiduroscopic lumbar discectomy through the sacral hiatus for sciatica is limited in quantity and quality. Therefore, this procedure should only be used in the context of research.

National Institute for Health and Care Excellence (United Kingdom) Percutaneous intradiscal laser ablation in the lumbar spine. Interventional procedures guidance[IPG357] (2010)64 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous intradiscal laser ablation in the lumbar spine is adequate to support the use of this procedure provided that normal arrangements are in place for clinical governance, consent and audit. Patients selected for the procedure should be limited to those with severe pain refractory to conservative treatment, in whom imaging studies show bulging of an intact disc, and who do not have neurological deficit requiring surgical decompression.

National Institute for Health and Care Excellence (United Kingdom) Automated percutaneous mechanical lumbar discectomy: Interventional procedures guidance[IPG141])(2005)65 Quality Rating: 2 out of 7

Current evidence suggests that there are no major safety concerns associated with automated percutaneous mechanical lumbar discectomy. There is limited evidence of efficacy based on uncontrolled case series of heterogeneous groups of patients, but evidence from small randomized controlled trials shows conflicting results. In view of the uncertainties about the efficacy of the procedure, it should not be used without special arrangements for consent and for audit or research.

(continued)



Table E-16 Synopsis of clinical practice guidelines related to lumbar radiculopathy or herniated invertebral lumbar disc (continued)



National Institute for Health and Care Excellence (United Kingdom) Endoscopic laser foraminoplasty. Interventional procedures guidance[IPG31] (2003)66 Quality Rating: 2 out of 7

Current evidence of the safety and efficacy of endoscopic laser foraminoplasty does not appear adequate to support the use of this procedure without special arrangements for consent and for audit or research.

a We assessed the quality of guideline using the Appraisal of Guidelines For Research & Evaluation II (AGREE II) Instrument,

version 2017.20 The lowest quality score possible is 1, the highest possible quality score is 7.

b Only recommendations from the guideline pertinent to surgical interventions for lumbar radiculopathy are summarized; see the

Full Report for a more complete summary.

c The complete guideline is not publicly accessible; thus, a full quality appraisal and summary of the evidence base and strength

of evidence ratings were not possible.

ES-4. Discussion

ES-4.1 Summary of the Evidence

Evidence maps summarizing the overall findings and strength of evidence are provided in

Figures 3, 4, 5, 6, and 7 of the Full Report. With few exceptions, most findings were based on

evidence graded as low to very low certainty.

Surgery reduces pain and improves function more than conservative management in the short

and medium term (up to 26 weeks) but this difference does not persist in the long term. Surgery-

related complications are rare and surgery may be cost-effective depending on a decision-makers

willingness to pay threshold.

With few exceptions, minimally-invasive surgery and standard surgery similarly reduce pain and

improve function. However, minimally-invasive surgery results in a quicker return to work and

lower incidence of persistent opioid use compared to standard surgery. Surgical morbidity and

reoperations are similar between both approaches. The evidence on cost-effectiveness for

minimally invasive surgery compared to standard approaches is inconclusive.

Microdiscectomy and discectomy are comparable with respect to efficacy and safety, but

microdiscectomy costs are higher.

For persons with recurrent lumbar radiculopathy, efficacy and safety outcomes are similar for

repeat lumbosacral decompression compared with spinal cord stimulation, but persistent opioid

use is less with spinal cord stimulation. In contrast, surgical morbidity, persistent opioid use and

return to work outcomes are more favorable with revision endoscopic discectomy compared to

standard microdiscectomy.

ES-4.2 Limitations of the Evidence Base

The studies we identified for inclusion in this HTA includes several limitations listed below.

Please refer to the Full Report for a more detailed description of each of these limitations.



Nearly half of included studies were high risk for bias.

Studies were generally underpowered to detect between-group differences, particularly

among studies comparing one minimally-invasive surgery to standard surgery.

Variation in diagnosis of radiculopathy and severity of symptoms among participants at

baseline.

Limited number of comparative effectiveness trials for any given procedure and lack of

complete surgical intervention description.

Variation in type, timing, and completeness in reporting outcomes.

All but two cost studies were conducted outside of the United States.

ES-4.3 Other Related HTAs

The only related HTA that we identified was commissioned by the National Institute for Health

Research (U.K.) Health Technology Assessment programme.67 This HTA included both surgical

and nonsurgical interventions for the management of sciatica and used a network meta-analysis

to provide a measure of relative therapeutic effect across 18 different treatment categories. The

findings suggest that nonopioid medication, epidural corticosteroids injections, and disc surgery

are effective for reducing sciatica. This HTA also concluded that stepped care approaches to

treatment are cost-effective relative to direct referral for surgery.

ES-4.4 Payer Coverage

The Centers for Medicare and Medicaid Services (CMS) does not have a national coverage

determination related to open standard or microsurgical decompressive procedures (i.e.,

discectomy, microdiscectomy, foraminotomy, laminectomy/otomy). CMS recognizes the use of

lasers to alter, revise, or destroy tissues in place of more conventional techniques as part of

surgical procedures.68 Medicare administrative contractors have been advised to use discretion in

determining coverage for procedures performed with a laser when the laser has been FDA-

approved, the procedure is considered reasonable and necessary, and a noncoverage instruction

does not exist (effective date May 1, 1997).68 CMS does have a national coverage determination

related to thermal intradiscal procedures; these procedures are not covered (effective date

January 1, 2009).68 Percutaneous disc decompression falls within the category of procedures

covered by this determination. Table ES-17 provides an overview of other payer coverage

policies; please see the Full Report for complete details.

In general, payers cover decompressive procedures, including discectomy, laminectomy/otomy,

foraminectomy/otomy, including microsurgical approaches, for disc herniation with radicular

symptoms. Specific criteria vary by payer but often include a failed trial of conservative

management for 6 to 12 weeks. Most payers also require imaging confirmation of nerve root

compression that corresponds to symptoms and physical examination findings. The coverage of

minimally invasive procedures varies by payer.



Table ES-17. Overview of payer coverage policies

Procedure Medicare Premera Regence Cigna United Aetna Humana Kaiser

Laminectomy, laminotomy, discectomy, foraminotomy (open technique including microsurgical approaches)

— a — — — a a —

Automated percutaneous lumbar disc decompression

b c —

(Percutaneous) endoscopic discectomy

No additional reimbursement.

—

(Percutaneous) laser discectomy


—

Percutaneous nucleoplasty with coblation technology

— —

Notes: = covered; = not covered; — = no policy identified

a If specific clinical criteria are met. See Table 45 in Full Report for details.

b All percutaneous disc decompression procedures fall under a Medicare National Coverage Determination related to thermal

intradiscal procedures.

c Also covers percutaneous manual discectomy, see Table 45 for details.

ES-4.5 Limitations of this HTA

This HTA was limited to studies and other information published or publicly available in

English. The electronic search was limited to only three databases. For efficiency, we relied on

hand searches of existing systematic reviews to identify eligible studies published prior to 2007.

Although this approach may have resulted in missed studies, we think this is unlikely since we

hand searched more than 40 systematic reviews. We used a single reviewer to screen titles and

abstract; however, we mitigated this risk through reviewer training, quantitative assessment of

interrater reliability during initial dual-review of 50 titles/abstracts, and by using a low threshold

for reviewers to request a second screening by another team member. We only included efficacy

outcomes reported at 4 weeks or later; thus, immediate and very-short term benefits are not

reflected in our synthesis. Lastly, the systematic review portion of this HTA was limited to trials;

observational evidence (i.e., case series and comparative cohort studies) were not considered.

ES-4.6 Ongoing Research and Future Research Needs

We did not identify any ongoing trials of surgical interventions specifically for lumbar

radiculopathy through our search of the U.S. clinical trials registry. Several trials related to

injections of biologics (e.g., condoliase into nucleus pulposus)69 or pharmacologics (e.g.,

epidural clonidine)70 or use of adjunctive treatments (e.g., epidural steroid injections, stem cell

injections, annular repair technologies) during or after discectomy to improve outcomes are

ongoing. The challenges faced in conducting methodologically rigorous randomized trials of

surgical interventions are well-documented.71 However, additional trials on treatment of lumbar

radiculopathy with the same methodologic flaws will be unlikely to change the certainty of

findings. Additional research on patient preferences and values related to timing of treatment or

surgery, and establishment of minimally important clinical differences in outcomes that are

specific to sciatica would advance research in this area. Finally, advanced analytic and statistical

techniques could be used within trials to quantify and mitigate the impact of crossovers on

treatment effects and could be used within observational studies to mitigate biases introduced by



nonrandomized study designs, potentially broadening the evidence base available to address

important research questions.

ES-5. Conclusion

Surgery for symptomatic lumbar radiculopathy reduced pain and improved physical function

more in the short and medium term (up to 26 weeks) compared to non-surgical interventions, but

these findings did not persist at one year or longer followup. Surgery compared with non-

surgical interventions resulted in similar improvements in neurologic symptoms, quality of life,

and return to work. No surgical deaths were reported by included studies and no differences in

persistent opioid use or all-cause mortality were observed. Surgery may be cost-effective relative

to non-surgical interventions depending on a decision maker’s willingness to pay threshold. All

findings were based on evidence graded as low to very low certainty.

Minimally invasive surgery was comparable to microdiscectomy or discectomy for reducing pain

and improving function in the short and long term based on most outcome measures reported.

Surgical morbidity and reoperations were similar in most studies, but a few studies suggested a

lower frequency of reoperation among participants that received standard surgery. Persistent

opioid use was lower among participants that received minimally-invasive surgery. The evidence

on cost-effectiveness of minimally invasive surgery is inconclusive. All findings were based


Microdiscectomy and discectomy were comparable with respect to efficacy and safety, but

microdiscectomy costs were higher. Efficacy and safety findings were based on evidence graded

as low to very low certainty; cost-effectiveness was based on evidence graded as moderate

certainty.

For persons with recurrent lumbar radiculopathy, efficacy and safety outcomes were similar for


use was less with spinal cord stimulation. In contrast, persistent opioid use and return to work

outcomes were more favorable with revision endoscopic discectomy compared to standard

microdiscectomy. All but one finding (surgical mortality) was based on evidence graded as very

low certainty.


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page -1

Full Technical Report

Structured Abstract

Purpose: To conduct a health technology assessment (HTA) on the efficacy, safety, and cost of

surgery for the treatment of symptomatic lumbar radiculopathy, also referred to as sciatica.

Data Sources: PubMed from January 2007 through November 9, 2017; clinical trial registry;

government, payor, and clinical specialty organization websites; bibliographies of relevant

clinical practice guidelines and systematic reviews.

Study Selection: Using a priori criteria, we selected English-language primary research studies

published in any year that were conducted in highly developed countries that enrolled adults with

symptomatic lumbar radiculopathy and compared surgery for radiculopathy to nonsurgical

interventions or that compared alternative surgical procedures. We selected trials or cost analyses

that reported efficacy outcomes (pain, functioning and disability, quality of life, neurological

symptoms, return to work) safety (mortality, surgical morbidity, reoperations, persistent opioid

use), or cost per quality-adjusted life year. We also selected the most recently published relevant

clinical practice guidelines (CPG) for quality appraisal and synthesis.

Data Extraction: One research team member extracted data and a second checked for accuracy.

Two investigators independently assessed risk of bias of included primary research studies and

conducted a quality assessment of included CPGs.

Data Synthesis: We included 25 primary research studies published between 1983 and 2017.

Twenty-four randomized controlled trials (RCTs) provided findings related to efficacy and

safety, 7 cost analyses provided findings related to cost-effectiveness. One RCT was rated as low

risk of bias, 12 were rated as having some concerns for bias, and 12 were rated as high risk of

bias. We identified 14 clinical practice guidelines.

Microdiscectomy and discectomy reduced leg pain by 6 to 26 points more than non-surgical

interventions as measured on a 0 to 100 point visual analogue scale of patient-reported pain at up

to 26 weeks followup; these findings did not persist at followup 1 year or later. Similar findings

were observed for functioning and disability as measured by the Oswestry Disability Index,

Roland-Morris Disability Questionnaire, and SF-36 Physical Functioning subscale. Surgery and

non-surgical interventions produced similar improvements in quality of life, neurologic

symptoms, and return to work. No surgical deaths occurred in any study and surgical morbidity

was infrequent. The incidence of reoperations among participants who underwent surgery ranged

from 0% to 10%; persistent opioid use was similar between groups. Studies reported higher

quality-adjusted life years for participants who underwent surgery compared to non-surgical

interventions, but similar or higher costs. The average cost per quality-adjusted life year gained

from a health care payor perspective ranged from $51,156 to $83,322 in 2010 US dollars.

Compared to microdiscectomy or discectomy, minimally-invasive surgery produced similar

improvements in pain, function/disability, quality of life, and neurologic symptoms, but resulted

in return to work 4 to 15 weeks sooner. No surgical deaths occurred in any studies and with few


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page -2

exceptions, surgical morbidity was similar. The incidence of reoperations across study groups

ranged from 2% to 64.5%; 2 of the 10 relevant studies reported much higher incidence of

reoperation among participants who underwent minimally-invasive procedures (64.5% vs. 2.5%

in 1 RCT; 44% vs. 16% in the other RCT) but the other 8 RCTs reported a similar incidence.

Compared to discectomy, microdiscectomy results in similar improvements pain, function and

other efficacy outcomes, and similar incidence of surgical morbidity and reoperations.

The four higher quality clinical practice guidelines we identified were in general agreement

about recommending discectomy or microdiscectomy (and related decompressive procedures) as

acceptable treatment based on evidence that it improves outcomes in the short to medium term.

Limitations: The included RCTs were limited by methodologic designs that increased risk for

bias, including extensive participant crossover, lack participant and outcome assessor blinding,

and inadequate randomization and allocation concealment in some studies. Many RCTs were

underpowered for outcomes of interest for this review, leading to imprecision for many effect

estimates reported. This HTA was limited to English-language studies reporting intent-to-treat

analyses; it did not include observational studies or ‘as-treated’ analyses reported by some RCTs.

Conclusions: Compared with non-surgical interventions, surgery reduces pain and improves

function more up to 26 weeks followup, but this difference does not persist at 1 year or longer.

Minimally invasive surgery, microdiscectomy, and discectomy are generally comparable with

respect to efficacy and surgical morbidity; findings were mixed for reoperations. Surgery may be

cost-effective when compared with non-surgical interventions, depending on a decision maker’s

willingness to pay threshold, but the evidence is inconclusive about the cost-effectiveness of

minimally invasive surgery. Nearly all findings are based on evidence with low to very low

certainty.


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page 1

1. Background

1.1 Purpose

To conduct a health technology assessment (HTA) on the efficacy, safety, and cost-effectiveness

of surgical interventions used to treat symptomatic lumbar radiculopathy, also known as sciatica

The State of Washington’s Health Technology Clinical Committee will use findings from this

assessment to inform policy decisions about surgery for radiculopathy.

1.2 Condition Description

Radiculopathy is a clinical syndrome characterized by radiating pain, with or without motor

weakness, and sensory disturbances in a myotomal or dermatomal distribution. When radicular

symptoms occur in the legs, this condition is referred to as lumbar radiculopathy or sciatica.

Nerve root compression is a common cause of radiculopathy and various pathological processes

may be responsible, but most often it results from disc herniation or spondylosis (i.e.,

degenerative joint and disc disease).3,72,73 Both processes can cause stenosis of the lateral

recesses or neural foramina and resulting spinal nerve root compression.3,72,73 Degenerative

changes can also produce spondylolisthesis, central spinal canal stenosis, and facet joint

hypertrophy, which may be associated with radiculopathy and nonradicular low back pain.3 Less

common etiologies of radiculopathy include infection, inflammation, neoplasm, vascular disease,

and congenital abnormalities.3,72 Radiculopathy is a clinical diagnosis because spinal nerve root

compression identified with imaging may not always be symptomatic. Thus, correlation of

symptoms and physical exam with imaging is usually used to diagnose radiculopathy, with

electromyography reserved for selected patients.

Lumbar radiculopathy is a heterogenous condition that may present acutely (as in the case of an

acute disc herniation) or more insidiously (as in the case of spondylosis).72 Further,

radiculopathy may present only with pain or with varying degrees of sensory disturbance or

motor weakness.74 Although stenosis of the central spinal canal more commonly presents with

neurogenic claudication, it can present with radicular symptoms.73

The objective of treatment for radiculopathy is symptom relief through nonsurgical management

of symptoms, or surgical intervention to address the underlying causative mechanism, or both.

Clinical trials of surgery to address causative mechanisms often specify persistent pain after 6

weeks of conservative management (i.e., medications, physical therapy, epidural steroid

injections, etc.) as a patient eligibility criterion for enrollment. In one observational study of 338

patients referred by general practitioners for low back pain or sciatica, 36% had major

improvements in symptoms within 2 weeks and 73% had reasonable to major improvements

within 12 weeks.75 A recent systematic review of preoperative predictors for postoperative

clinical outcomes following lumbar surgery for disc herniation reported 17 predictors of a

favorable outcome, 20 predictors of a nonfavorable outcome, and 15 predictors with conflicting

evidence of impact on outcomes.76

1.3 Disease Burden

Estimates of the prevalence of lumbar radiculopathy vary widely, likely due to variation in

definitions and differences between self-reported and clinically assessed symptoms.1 Several



studies report the lifetime prevalence of lumbar radiculopathy ranges from approximately 3% to

43%.1-3 Among three studies (two conducted in Finland and one in the United Kingdom (U.K.)

that defined sciatica by clinical assessment, the lifetime prevalence was 12.2%, the point

prevalence was 4.8%, and the period prevalence over 4 to 52 weeks ranged from 5.1% to 6.3%

within the general population.

Although some studies report that radiculopathy is distributed equally between men and women,4

others have found that men are more commonly affected.3,5 The highest prevalence likely occurs

between the ages of 45 and 64 years.2 Men may be more likely to develop symptoms in their

forties while women are at a higher risk in their fifties.3 Previous history of axial low back pain

is an established risk factor for radiculopathy.4,5 Other risk factors include a prior history of

trauma, prolonged driving, pregnancy, and jobs requiring manual labor.2,4,6 Several studies have

shown a genetic linkage for spinal canal size as well as disc herniation.4,7

1.4 Technology Description

Table 1 provides descriptions of the surgical procedures used to treat lumbar radiculopathy.

Standard, open surgical interventions remove parts of the intervertebral disc, with or without

additional decompression of spinal nerve root(s) through removal of parts of the bony vertebrae,

facet joints (e.g., laminectomy or partial facetectomy) and/or other soft tissues impinging on the

nerve root(s). Decompression and disc removal interventions are often performed with a

microscope or other magnifying instrument (“micro” approaches). Such an approach makes it

possible to minimize the length of incision and area of dissection, thereby reducing the degree of

structural alteration to surrounding tissues. Both standard open and “micro” approaches allow for

direct visualization of the disc and surrounding structures. When spondylolisthesis is present

preoperatively, or when spinal instability resulting from decompressive or disc removal

procedures is likely (e.g., the surgery involves vertebrae at higher lumbar levels or the nerve root

compression is in the cephalocaudal direction), then arthrodesis, (i.e., spinal fusion) may also be

required. In contrast, minimally invasive surgical interventions use either an endoscopic

approach to allow direct visualization of the surgical field and anatomy, or use a percutaneous

approach that does not allow direct visualization of the disc and surrounding tissue.

Table 1. Description of surgical interventions used to treat lumbar radiculopathy

Decompression procedures

Laminectomy is the removal of part or all of the lamina portion of the vertebra bone and ligaments to reduce compression.

Microlaminectomy is the removal of part or all of the lamina portion of the vertebra bone, but using a smaller incision and minimal dissection and disruption to surrounding bones, joints, ligaments, and muscles.

Laminotomy is the creation of an opening in the lamina portion of the vertebra bone to reduce compression.

Foraminotomy is the creation of a wider neuroforaminal opening of the vertebra to reduce compression within the neural foramina.

Disc removal procedures

Discectomy is the removal of some or all the intervertebral disk to reduce compression, and is sometimes combined with laminotomy or laminectomy or other procedures.

Microdiscectomy is removal of some or all of the intervertebral disk using a smaller incision and with minimal disruption of the surrounding bones, joints, ligaments, and muscles.

Other interventions that do not structurally alter the spine, often referred to as minimally invasive procedures



Percutaneous and/or endoscopic discectomy, discoplasty, nucleotomy, or nucleoplasty use mechanical (manual or automated), radiofrequency thermal, coblation (also known as plasma), or laser-assisted techniques for disc removal, destruction, or decompression.

Chemonucleolysis is the injection of enzymes (e.g., chymopapain) or other chemical substances into the nucleus pulposus to induce disc shrinkage and reduce compression. This procedure is no longer in routine clinical use.

1.5 Regulatory Status

Surgical procedures are not regulated; however, the U.S. Food and Drug Administration (FDA)

(through the 510(k) process) may regulate surgical instruments and devices used in surgery.

FDA-approved electrosurgical cutting instruments, aspiration or coagulation devices,

endoscopes, and other related accessories can be found under product codes GEI, GXI, HRX,

BSO, and BSP. The following summary is not an all-inclusive list.

The FDA has approved several devices for aspiration during percutaneous discectomies

including the DeKompressor® Percutaneous Discectomy Probe77 (Stryker, approved November

7, 2003), Herniatome Percutaneous Discectomy Device78 (Gallini Medical Devices, approved

December 8, 2014), Nucleotome®79 (Clarus Medical, approved June 1, 2004), and Laurimed

Percutaneous Discectomy System80 (Laurimed LLC, approved August 28, 2008). These are all

Class II devices indicated for “aspiration of disc material during percutaneous discectomies in

the lumbar, thoracic and cervical regions of the spine.” The SpineView ENSPIRET™81 Debrider

(SpineView Inc, April 15, 2009) and the enSpire™ Discectomy System82 (Spine View Inc,

approved June 26, 2012) are Class II devices indicated for use in “cutting, grinding and

aspirating intervertebral disc material during discectomy procedures in the cervical, thoracic and

lumbar spine.”

Numerous devices have received approval for ablation and coagulation. A commonly noted

device in several payer coverage policies is the Disc-FX™ system83 (Ellman International Inc,

approved February 27, 2006), a Class II device indicated for “ablation and coagulation of

intervertebral disc material during discectomy procedures in the lumbar spine.” Several Class II

catheters, such as the SpineCATH Intradiscal Catheter84 (ORATEC Interventions Inc, approved

December 17, 1999) and the Nucleotomy Catheter85 (ORATEC Interventions Inc, approved

January 31, 2002), have been approved for the indicated use of “coagulation and decompression

of disc material.” One included study in this HTA39 used the Perc-D SpineWand™86 (Arthrocare,

approved August 22, 2013), a Class II device indicated for “ablation, coagulation, and

decompression of disc material,” to conduct percutaneous disc decompression with coblation

technology.

One study included in this HTA27 utilized the Metr’X system arthroscope (Medtronic Sofamor

Danek, Inc., approved November 24, 2000).87 Frequently cited in insurance coverage policies,

the Yeung Endoscopic Spine System (Richard Wolf Instrument Company, approved March 13,

1998)88 is a Class II device indicated for the “visualization and removal of herniated discs in the

lumbar region.” Though also often discussed in payer policies as a medical device, the

AccuraScope™ Discectomy and Neural Decompression (DND) procedure has been marketed by

North American Spine, a private provider of minimally invasive spine surgery,89 but is not FDA-

regulated.

FDA has approved laser instruments for incision, excision, resection, ablation, vaporization, and

coagulation of tissue during surgical procedures including but not limited to discectomy. One



included study in this HTA37 utilized the Ceralas 980 Diode Laser90 (Biolitec Inc, approved

January 24, 2008), a Class II device indicated for “delivery of laser light to soft tissue in the

contact or non-contact mode during surgical procedures, including via endoscopes, introducers,

or catheters,” to conduct percutaneous laser disc decompression. For more information on

coverage of surgical procedures that use laser devices, see the Selected Payer Coverage Policies

section of this report.

1.6 Policy Context

Numerous surgical and nonsurgical approaches to the management of lumbar radiculopathy are

routinely used within current clinical practice. In addition to standard surgical techniques (e.g.,

laminectomy, discectomy), minimally invasive surgical techniques that use percutaneous,

endoscopic, or laser-assisted approaches are now available. The State of Washington Health Care

Authority selected surgery for lumbar radiculopathy as a topic for an HTA based on medium

concerns for efficacy, medium concerns for safety, and high concerns for cost. This HTA is

designed to assist the State of Washington’s Health Technology Clinical Committee in

determining coverage for selected surgical interventions to treat symptomatic lumbar

radiculopathy.

1.7 Washington State Agency Utilization Data

The State of Washington Health Care Authority provided data on the use of various surgical

procedures for the treatment of radiculopathy for the time period X to X.

2. Methods

This HTA includes two separate, but related components. The first component is a systematic

review of primary research studies and the second component is a quality appraisal and synthesis

of relevant clinical practice guidelines.

2.1 Research Questions and Analytic Framework for Systematic Review

of Primary Research Studies

We developed the following research questions and analytic framework (Figure 1) to guide the

systematic evidence review of primary research studies:

Efficacy Question 1 (EQ1). In adults with symptomatic lumbar radiculopathy, what is the

effectiveness and comparative effectiveness of surgical interventions?

Efficacy Question 2 (EQ2). In adults with symptomatic lumbar radiculopathy, does

effectiveness or comparative effectiveness of surgical interventions vary for patients who are not

employed because of disability or patients who are undergoing recurrent surgery for relapse?

Safety Question 1 (SQ1). In adults with symptomatic lumbar radiculopathy, what are the

adverse events associated with surgical interventions?

Cost Question 1 (CQ1). In adults with symptomatic lumbar radiculopathy, what is the cost-




Figure 1. Analytic framework for HTA on surgery for lumbar radiculopathy

Adults with

symptomatic

lumbar

radiculopathy

Surgical

interventions to

reduce pain,

symptoms, and

improve function

Pain



Physical, psychological, and

social functioning

Return to work

Reoperations for relapse/

recurrent symptoms

$/quality-adjusted life year

gained

$/disability-adjusted life

year/gained

Surgery-related morbidity and mortality



EQ1

SQ1

CQ1

EQ2

Abbreviations: CQ=cost question; DD=developmental disability; EQ=efficacy question; SQ=safety question

The State of Washington Health Technology Assessment Program posted a draft of these

research questions and proposed study selection criteria for public comment from November 14,

2017 to November 27, 2017. A summary and response to public comments received is available

at the Program’s website.91

2.1.1 Data Sources and Searches

We searched MEDLINE® (via PubMed) from 2007, the Cochrane Library, and a clinical trials

registry (clinicaltrials.gov) for relevant English-language studies. We searched the CMS and

FDA websites, selected payer and health care professional society websites, and websites of

other organizations that conduct and disseminate HTAs. In addition, we reviewed the reference

lists of relevant studies, practice guidelines, and other HTAs on this topic to identify any relevant

articles not found through the electronic search and to identify studies published prior to 2007.

The detailed search strategy is provided in Appendix A.

In brief, we used medical subject headings (MeSH terms) and text words associated with the

surgical interventions of interest combined with MeSH terms for radiculopathy and lumbar disc

disease. We limited the search by eliminating studies indexed using terms for infants, children,

or adolescents, selected animals, and conditions that indicate excluded populations, such as

cancer, tuberculosis, fracture, scoliosis, spondylolysis, and cervical vertebrae. We used MeSH

terms to select studies most likely to be trials or systematic reviews and to remove editorials,

letters, and publication types that do not represent primary research studies.

2.1.2 Study Selection

Table 2 summarizes the final study selection criteria related to the population, intervention,

comparator, outcomes, time period, and setting that defined the scope of this HTA; these are



further described after the table. We screened titles and abstracts and full-text articles based on

these study selection criteria. Four team members independently screened the same initial 50

titles/abstracts. Because we had substantial interrater reliability (Light’s Kappa = 0.84),92 the

remaining titles/abstracts were divided among team members and a single reviewer screened

them. The lead investigator and one additional team member independently screened all full-text

articles; discrepancies were resolved by discussion.

Table 2. Population, intervention, comparator, outcome, timing, setting and other study selection criteria for HTA on surgery for lumbar radiculopathy


Population Adults age 18 years and over with symptomatic lumbar radiculopathy (i.e., sciatica) unrelated to infection, cancer, inflammatory, congenital, or traumatic etiologies. For studies of mixed populations, results must be stratified and reported separately for patients with lumbar radiculopathy.

Adults with:

Cervical or thoracic radiculopathy

Cauda equina syndrome

Neurogenic claudication or low back and leg symptoms related primarily to central spinal stenosis

Spondylolisthesis

Traumatic or congenital structural spinal abnormalities

Nonradicular leg or low back pain (i.e., discogenic or other nonspecific low back pain)

Intervention Surgical interventions for the treatment of radiculopathy, for example:

Laminectomy, laminotomy

Discectomy

Foraminotomy

Nucleotomy Includes “micro” approaches to the above procedures, which involve smaller incisions and/or areas of dissection and/or use of microscope or loupe magnification. Minimally invasive surgical procedures designed for treating radicular pain: percutaneous discectomy, discoplasty, nucleotomy, or nucleoplasty that are manual, automated, endoscopic, or laser-assisted, or use radiofrequency heat, including coblation technology.

Interventions involving combinations of the above interventions are eligible.

Surgical interventions primarily designed to treat neurogenic claudication and central spinal stenosis, spinal instability, or nonradicular low back pain, for example:

Spinal fusion

Arthroplasty

Artificial disc replacement

Interspinous process decompression (e.g., X-STOP® IPD System,8 Coflex® Interlaminar Technology)9

Minimally invasive lumbar decompression (mild® procedure)10

Other minimally invasive procedures designed for treating discogenic (i.e., nonradicular) low back pain

Epidural, spinal, or disc injections of enzymatic (e.g., chymopapain), chemical, or biologic (e.g., stem cells, mesenchymal cells) agents. Interventions involving combinations of procedures that include an above intervention are ineligible.

(continued)



Table 2. Proposed population, intervention, comparator, outcome, timing, and setting for HTA on surgical interventions for lumbar radiculopathy (continued)


Comparator Placebo or no treatment comparators: sham surgery, expectant management, no treatment Active treatment comparators:

Nonsurgical management (e.g., physical therapy, exercise, pharmacologic treatment of symptoms, spinal manipulation, chiropractic treatment, epidural steroid or pain injections, other noninvasive treatments)

Surgical interventions as listed under “intervention”

No comparator

Chemonucleolysis Studies using “usual care” comparator groups will not be excluded but will be synthesized separately if no information was provided about the components of “usual care.”

Outcomes Efficacy (at 4 weeks post-op or later):

Pain

Physical functioning

Social functioning

Psychological/emotional distress


Neurologic symptoms (e.g., weakness, sensory alteration)

Return to work

Reoperations for relapses/recurrent symptoms

Safety:

Surgery-related morbidity (e.g., venous thromboembolism, paralysis, new neurological symptoms, dural tear, epidural hematoma)

Surgical mortality (30 day)


Persistent opioid use Cost-effectiveness: Cost per quality-adjusted life year gained Cost per disability-adjusted life year gained

Other outcomes not specifically listed as eligible. Pain, quality of life, and functional outcomes not measured using valid and reliable instruments or scales.11,12

Setting Inpatient or outpatient settings in countries categorized as “very high” on United Nations Human Development Index

Studies conducted in countries not categorized as “very high” on United Nations Human Development index.

Study Design and Risk of Bias Rating

For all Efficacy and Safety Research Questions: CCTs, RCTs, and SRs of CCTs or RCTs with similar scope as this HTA. For studies using active comparators, only RCTs or SRs of RCTs will be included. For Cost-Effectiveness Questions: CEA, CUA, or CBA performed from the societal or payer perspective For all studies: Any risk of bias rating, but high risk of bias studies will only be used in quantitative syntheses if fewer than 3 studies total are available.

Editorials, comments, letters, narrative reviews, case reports, case series, cohort studies, case-control studies.

(continued)



Table 2. Proposed population, intervention, comparator, outcome, timing, and setting for HTA on surgical interventions for lumbar radiculopathy (continued)


Language and Time Period

English, no restrictions on time period included. Languages other than English.

Abbreviations: CCT = controlled clinical trial; CEA = cost-effectiveness analysis; CUA = cost-utility analysis; CBA = cost-

benefit analysis; HTA = health technology assessment; RCT = randomized controlled trial; SR = systematic review

2.1.2.1 Population

Studies were selected if they enrolled adults age 18 years or over with symptomatic lumbar

radiculopathy unrelated to infection, cancer, inflammatory, congenital or traumatic etiologies.

Studies that enrolled participants with lumbar spinal stenosis and neurogenic claudication,

spondylolisthesis, or chronic discogenic (i.e., nonradicular) low back pain were excluded.

2.1.2.2 Intervention and Comparator

For all research questions, comparative studies where at least one study group included a surgical

intervention to relieve lumbar radicular pain were eligible for selection. This included standard

open discectomy with or without laminotomy and laminectomy, foraminotomy, sequestrectomy,

or nucleotomy. Any forms of these surgeries including minimally invasive approaches (e.g.,

endoscopic or percutaneous approaches), use of microsurgical techniques (e.g.,

microdiscectomy), or laser-assisted procedures were eligible. Comparison groups that were

placebo or no treatment comparators or active treatment comparators were eligible for selection.

Active treatment comparators could include nonsurgical management (e.g., analgesics, physical

therapy, spinal manipulation, epidural injection, etc.) or could include an alternative surgical

intervention (e.g., endoscopic discectomy compared with standard open discectomy).

2.1.2.3 Outcomes

For the research questions on efficacy (EQ1, EQ2), studies that reported outcomes related to

pain, quality of life, and functional outcomes were eligible for selection and we generally

required studies to use valid and reliable measures of these constructs (e.g., SF-36, Roland-

Morris Disability Questionnaire, etc.). In addition, studies that reported on change in neurologic

symptoms and return to work were also eligible for selection. For the research question on safety

(SQ1), studies that reported on perioperative or postoperative morbidity and mortality,

reoperations, or persistent opioid use were eligible for selection. For the research question on

cost (CQ1), studies that reported cost-effectiveness measures, specifically cost per quality-

adjusted life year gained (QALY) or cost per disability-adjusted life year gained (DALY) were

eligible for selection.

2.1.2.4 Settings

Studies conducted in any inpatient or outpatient clinical setting were eligible for selection.

Studies that were conducted in countries designated as “very high” by the UN Human

Development Programme were eligible for selection as these countries (Europe, Australia, New

Zealand, Japan, S. Korea, Singapore, Hong Kong, and selected Middle Eastern countries) are

like the United States. We excluded studies conducted in countries designated as less than “very

high.”13



2.1.2.5 Study Design

Nonrandomized controlled clinical trials, randomized controlled trials (RCTs), and systematic

reviews of trials were eligible for selection. However, we required trials comparing two

alternative surgeries to be randomized. Cohort studies, studies that analyzed randomized trial

data with cohort study data together, case series, and case reports were not eligible for selection.

2.1.2.6 Time Period

We did not restrict included studies based on year conducted or published.

2.1.3 What is Excluded from This HTA

This review will not include studies published in languages other than English on conducted in

countries that are considered not very highly developed (based on UN Human Development

Programme).13 We also exclude studies evaluating surgical interventions performed primarily to

manage symptoms of central spinal canal stenosis (e.g., neurogenic claudication),

spondylolisthesis, traumatic or congenital abnormalities, or radiculopathy resulting from

infectious or neoplastic processes. This review will not cover surgical interventions for chronic

discogenic low back pain that is not radicular. This review will also not include studies

conducted among children or adolescents.

2.1.4 Data Abstraction and Risk of Bias Assessment

One team member extracted relevant study data into a structured abstraction form and the lead

investigator checked it for accuracy. For consistency in reporting findings across studies, we

transposed some treatment effects reported in studies to ensure all our abstracted data

represented the effect of the intervention group relative to the comparator group. We used the

Cochrane Risk of Bias (RoB 2.0) tool to assess the risk of bias for each included study.14

Domains assessed with this tool include: bias arising from randomization process, bias due to

deviations from intended interventions, bias due to missing outcome data, bias in measurement

of the outcome, and bias in selection of the reported result. Risk of bias was assessed as “high,”

“some concerns,” or “low” at the study level unless different outcomes within a single study

required different risk of bias ratings. Two team members conducted independent risk of bias

assessments on all included studies; discrepancies were resolved by discussion, in consultation

with the lead investigator if needed.

2.1.5 Data Synthesis and Strength of Evidence Rating

Study characteristics and results were qualitatively synthesized for each research question in

tabular and narrative formats. We synthesized studies comparing the surgical interventions to

nonsurgical interventions separately from studies comparing alternative surgical interventions.

We summarized continuous outcome measures as absolute mean differences (AMD) wherever

possible. When studies did not report the AMD for critical outcomes, we calculated it and the

95% confidence intervals (CI) if the appropriate data were reported in the article to be able to do

so (e.g., mean, standard deviation (SD) for each comparison group). We summarized

dichotomous outcomes using proportions, and when studies did not report tests of statistical

significance we calculated P values using Fisher’s exact test. We identify all values that we

calculated in the text and tables as “calculated” values. For cost outcomes, we adjusted all

reported outcomes in foreign currency to U.S. dollars based on the U.S. Department of Treasury



mid-year exchange rate for the year reported by study authors and then used the chain-weighted

consumer price index (CPI) to adjust to 2010 U.S. dollars (Appendix B).15,16

To determine whether quantitative synthesis was appropriate, we assessed the number of studies

and the clinical and methodological heterogeneity present based on established guidance.93,94 We

required three or more studies with similar intervention and comparator with same outcome

measure at approximately the same follow-up time point to calculate a pooled effect estimate.

We considered outcomes reported at less than 12 weeks to be short-term, outcomes reported

between 12 weeks up to 52 weeks as medium-term, and outcomes reported at 52 weeks or later

as long-term. We estimated pooled effects using a random effects model with the ‘metafor’

package in R using the DerSimonian and Laird method.17

We graded the strength of evidence for each research question and each outcome measure using

GRADE, which assesses the strength of evidence based on domains relating to risk of bias,

inconsistency, imprecision, indirectness, and other considerations, such as publication bias.18 For

each outcome measure, we rated the evidence for short-, medium-, and long-term outcomes

separately when required. With GRADE, the strength of evidence can be graded as “very low,”

“low,” “moderate,” or “high”, and this rating represents the overall certainty of the findings.

Evidence graded as “high” can be interpreted to mean that future studies are very unlikely to find

substantively different findings, whereas evidence graded as “very low” means that future

studies could have substantively different findings. An evidence base consisting of RCTs begins

with a high strength of evidence rating based on study design but can be downgraded to

moderate, low, or very low based on serious or very serious concerns in the domains assessed.

2.2 Clinical Practice Guideline Synthesis

In addition to the systematic evidence review portion of this HTA, we also identified relevant

clinical practice guidelines and conducted a quality assessment of each guideline using the

Appraisal of Guidelines for Research & Evaluation II instrument.19,20 With this instrument, six

domains are assessed and an overall score of between 1 (lowest possible) and 7 (highest

possible) are assigned to reflect the overall quality of the guideline.

3. Results

3.1 Literature Search

Figure 2 depicts the study flow diagram. We identified and screened 1,860 unique citations. We

excluded 1,638 citations after title and abstract review. We reviewed the full-text of 222 articles

and included a total of 25 studies reported in 38 articles published between 1983 and 2017.

Twenty-two RCTs provided evidence on efficacy or comparative effectiveness (EQ1), two RCTs

provided evidence on the effectiveness or comparative effectiveness of revision surgical

interventions for relapse (EQ2), 24 RCTs provided evidence on safety (SQ1), and seven studies

(six RCTs and one cost-effectiveness analysis) provided evidence on costs or cost-effectiveness

(CQ1).



Figure 2. Study flow diagram for HTA on surgery for lumbar radiculopathy

Number of citations identified through database searches:

1789

Number of titles/abstracts after duplicates removed:

1750

Number of titles/abstracts screened: 1860

Number of titles/abstracts excluded:

1638

Number of titles/abstracts

identified through hand searches:

110

Number of full-text articles assessed for eligibility:

222

Number of studies included: 25 studies from 38 articles

Number of full-text articles excluded: 184

By reason: Systematic review from handsearch: 42 Ineligible publication type: 15 Ineligible country: 15 Ineligible population: 13 Ineligible intervention: 6 Ineligible comparator: 13 Ineligible outcome: 7 Ineligible study design: 41 Duplicate or superseded: 13 Study protocol or in progress: 6 Abstract only: 6 Non-English full text: 6 Data uninterpretable: 1a

Number of studies included for EQ1 and EQ2 (Efficacy):

24

Number of studies included for SG1 (Safety):

24

Number of studies included for CQ1 (Cost):

7

aWe contacted the study author for clarification but did not receive a reply.



Individual study and population characteristics and findings for all included studies are

summarized in Appendix C. The list of studies we screened at the full-text stage, but which we

excluded, is provided in Appendix D. Note that studies may have been excluded based on more

than one reason but we report only one reason. We report our individual study risk of bias

assessments for included studies in Appendix E.

3.2 Efficacy

3.2.1 Efficacy Question 1

In adults with symptomatic lumbar radiculopathy, what is the effectiveness and comparative


We included 22 RCTs. Seven RCTs provided evidence for the efficacy of surgery compared with

nonsurgical treatment; study and population characteristics are summarized in Table

3.5,21,22,30,31,35,39 Fifteen RCTs (providing 17 comparisons) provided evidence for the comparative

effectiveness of alternative surgical interventions; study and population characteristics are

summarized in Table 4.23-29,33,34,36-38,40-42 The section that follows these tables describes study

characteristics and findings in detail along with a summary of findings table that includes

strength of evidence rating. Appendix C and Tables C-1 and C-2 provide evidence tables with

individual study and population characteristics. Appendix C and Tables C-3 and C-4 provide

evidence tables with detailed individual study outcomes related to efficacy.+

3.2.1.1 Study Characteristics

All 22 included studies were parallel-group RCTs. Four were conducted in the U.S.;21,39-41 the

rest were conducted in Canada (N=1),22 Taiwan (N=1),23 Japan (N=1)24 or various European

countries (N=15).5,25-38 The total number of participants randomized ranged from as few as 21 to

as many as 501. The mean age of participants generally ranged from mid-30s to mid-40s. All

studies enrolled both men and women with signs or symptoms of lumbar radiculopathy and

confirmatory imaging, usually CT or MRI. Most studies excluded participants with indications

for immediate surgery, for example, cauda equine syndrome. Only six studies reported the

proportion of participants disabled at baseline,21,22,25,30,37,38 and they did not use a consistent

definition of disability. The mean duration of symptoms at baseline ranged from 8 weeks to 2

years among the 13 included studies reporting this population characteristic.22,25-27,29-32,36-39,41



Table 3. Study and population characteristics of the seven randomized controlled trials comparing surgery to nonsurgical interventions for management of lumbar radiculopathy (EQ1)

Author (Year); Country; Risk of Bias

Population Characteristics Age, mean (SD); Women N (%); Duration of symptoms, mean (SD)

Surgical Intervention (SG1); N randomized; N analyzed (% of randomized); N crossovers (% of randomized)

Comparator(s) (NS1); N randomized; N analyzed (% of randomized); N crossovers (% of randomized)

Primary Outcome (effect size detectable with

80% power, = 5%)a;

Other outcomes

Erginousakis (2011)35; Greece; High

Age SG1: 38 (4.2); NS1: 36 (5.8) Women: SG1: 12 (38.7%); NS1: 14 (45.2%) Duration of symptoms: NR

Percutaneous disc decompression N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%)

Conservative management N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%)

VAS 10 Pain (NR);

N (%) with category of pain reduction

N (%) reporting pain affected occupational status

Gerszten (2010)39 United States; Some concerns (6w outcomes); High (12w and later outcomes)

Age SG1: 46 (12); NS1: 42 (11) Women SG1: 24 (53%); NS1: 19 (48%) Duration of symptoms: SG1: 52w (range 4w to 16y) NS1: 2y (range 10w to 13y)

Plasma disc decompression with coblation technology N randomized: 46; 45 ITT sample; N analyzed: 29 (64% of ITT sample) at 26w; N crossovers: 12 were unresolved and received a second, unspecified procedure.

Epidural steroid injection N randomized: 44; 40 ITT sample; N analyzed: 28 (70% of ITT sample) at 26w; N crossovers: 8 were unresolved and received a second, unspecified procedure.

VAS 100 leg pain (15 points);

VAS 100 back pain

SF-36 bodily pain

Oswestry disability index

SF-36 physical functioning

SF-36 Physical Component and Mental Component Summary

Other SF-36 subscales

N (%) with full muscle strength

N (%) with normal tactile sensitivity

N (%) working full or part-time

McMorland (2010)22 Canada; Some concerns

Age SG1 Men: 42.85 (NR); SG1 Women: 40.1 (NR); NS1 Men: 36.4 (NR); NS1 Women: 48.33 (NR) Women SG1: 7 (35%); NS1: 9 (45%) N (%) with duration of complaint 12-26w SG1: 3 (15%); NS1: 6 (30%)

Microdiscectomy N randomized: 20; N analyzed: 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w: 20 (100%) 52w: 15 (75%); N crossovers: 3 (15%) enrolled in spinal manipulation 26-34w after surgery so received both interventions.

Spinal manipulation N randomized: 20; N analyzed: 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w: 20 (100%) 52w: 17 (85%); N crossovers: 8 (40%) underwent microdiscectomy after 12w of spinal manipulation care.

Aberdeen Back Pain Scale (6 points);

McGill Pain Questionnaire

Roland-Morris Disability Questionnaire

SF-36 bodily pain



SF-36 Total Score

(continued)



Table 3. Study and population characteristics of the seven randomized controlled trials comparing surgery to nonsurgical interventions for management of lumbar radiculopathy (EQ1) (continued)






80% power, = 5%)a;

Other outcomes

Osterman (2003)31 Finland; High

Age SG1: 37 (7); NS1: 38 (7) Women; SG1: 13 (46.4%) NS1: 9 (32.1%) Duration of leg pain, mean (SD) SG1: 11.0w (4.6); NS1: 8.6w (3.0)

Microdiscectomy N randomized: 28; N analyzed 6w: 26 (93%) 26w: 26 (93%) 52w: 21 (75%) 2y: 26 (93%); N crossovers: 0 (0%)

Physiotherapy N randomized: 28; N analyzed 6w: 26 (93%) 26w: 22 (78.6%) 52w: 20 (71.4%) 2y: 24 (86%); N crossovers: 11 (39.3%)

VAS 100 leg pain (15 points);

VAS 100 back pain

VAS 100 Work Ability Score


15D HRQOL

N (%) with muscle weakness

N (%) reporting full recovery

Peul (2007)30 Peul (2008)95 Lequin (2013)96 Netherlands; High

Age SG1: 41.7 (9.9); NS1: 43.4 (9.6) Women SG1: 52 (37%); NS1: 45 (32%) Duration of symptoms, mean (SD) SG1: 9.43w (2.37w); NS1: 9.48w (2.11w)

Microdiscectomy N randomized: 141; N analyzed 52w: 140 (99.3%) 2y: 130 (92.2%) 5y: 115 (81.6%); N crossovers 52w: 16 (11.3%) 2y: 16 (11.3%) 5y: 16 (11.3%);

Conservative management N randomized: 142; N analyzed 52w: 141 (99.3%) 2y: 130 (91.5%) 5y: 116 (81.7%); N crossovers 52w: 55 (38.7%) 2y: 62 (43.7%) 5y: 66 (46.5%);

Roland-Morris Disability Questionnaire (90% power, 3 points);

VAS 100 leg pain

VAS 100 back pain

Sciatica index

SF-36 bodily pain



Prolo Scale

Median time to recovery

Global perception of recovery

Weber (1983)5 Norway; High

Age SG1: 40.0 (NR); NS1: 41.7 (NR) Women SG1: 28 (46.7%); NS1: 30 (45.5%) Duration of symptoms: NR

Discectomy N randomized: 60; N analyzed: 60 (100%); N crossovers: 1 (1.7%)

Conservative management N randomized: 66; N analyzed: 66 (100%); N crossovers: 17 (25.8%)

NR (NR);

N (%) with categories of radiating pain

N (%) with permanent incapacitation and receiving disablement benefits

N (%) with category of result (Good, Fair, Poor Bad)

(continued)



Table 3. Study and population characteristics of the seven randomized controlled trials comparing surgery to nonsurgical interventions for management of lumbar radiculopathy (EQ1) (continued)






80% power, = 5%)a;

Other outcomes

Weinstein (2006) 21 Weinstein (2008)45 Lurie Jon (2014)44 United States; High

Age SG1: 41.7 (11.8); NS1: 43.0 (11.3) Women SG1: 101 (44%); NS1: 93 (39%) Duration of symptoms: NR

Discectomy/microdiscectomy N randomized: 245; N analyzed: 232 in main study's primary analyses. N crossovers: Cumulative over time: 26w: 113 (46.1%) 2y: 105 (42.9%) 8y: 97 (39.6%)

Conservative management N randomized: 256; N analyzed: 240 included in main study's primary analyses. N crossovers: Cumulative over time: 26w: 93 36.3%) 2y: 107 (41.8% 8y: 122 (47.7%)

SF-36 bodily pain, SF-36 physical functioning, Oswestry disability index (85% power, 10 points in SF-36 scales or “similar effect size in the Oswestry disability index”);

Sciatica index

N (%) working full or part time

N (%) satisfied with symptoms

N (%) with self-rated major improvement

a As specified and reported by study authors.

Abbreviations: HRQOL = health-related quality of life; ITT = intent to treat; N = number; NR = not reported; NS= nonsurgical group; SD = standard deviation; SF-36 = short-

form 36 survey; SG = surgical group; VAS = visual analogue scale; w = week(s); y = years(s).



Table 4. Study and population characteristics of the 15 randomized controlled trials comparing alternative surgical interventions for management of lumbar radiculopathy (EQ1)



Surgical Intervention (SG1); N randomized; N analyzed (% of randomized); N crossovers (% of randomized);

Comparator(s) (SG2, SG3); N randomized; N analyzed (% of randomized); N crossovers (% of randomized);


80% power, = 5%)a;

Other outcomes

Minimally-invasive surgery compared with microdiscectomy or discectomy

Arts (2009)38 Arts (2011)43 The Netherlands; Low

Age SG1: 41.6 (9.8); SG2: 41.3 (11.7) Women SG1: 82 (49%); SG2: 71 (45%) Duration of symptoms, median (range) SG1: 21.0w (13 to 30) SG2: 21.0w (13 to 34)

Tubular discectomy N randomized: 167; N analyzed: 166 (99.4%); N crossovers: 2 (1.2%)

Microdiscectomy N randomized: 161; N analyzed: 159 (98.8%); N crossovers: 0 (0%)


VAS 100 leg pain

VAS 100 back pain

SF-36 bodily pain


Sciatica index

Prolo Scale


Brouwer (2015)37 Brouwer (2017) 97 The Netherlands; Some concerns

Age SG1: 43.2 (11.8); SG2: 43.7 (9.7) Women SG1: 19 (35%); SG2: 24 (42%) Duration of sciatica, median (range) SG1: 30.0w (9 to 182) SG2: 26.0w (8 to 260)

Percutaneous laser disc decompression N randomized: 57; N analyzed: 55 (96.5%); N crossovers: Unclear



VAS 100 leg pain

VAS 100 back pain

SF-36 bodily pain


Sciatica index

Prolo Scale (functional and economic subscales)


Chatterjee (1995)36 United Kingdom; Some concerns

Age SG1: 38.9 (range 20 to 56); SG2: 41.3 (range 21 to 67) Women SG1: NR (51%); SG2: NR (40%) Duration of current episode of radicular pain, mean (range) SG1: 13w (6 to 30) SG2: 20w (6 to 38)

Automated Percutaneous Lumbar Discectomy N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%) in main analysis, 20 eventually underwent microdiscectomy

Microdiscectomy N randomized: 40; N analyzed: 40 (100%); N crossovers: 0 (0%)

NR (NR);

N (%) returned to work or previous level of activity

N (%) with excellent/good outcome based on MacNab criteria

(continued)



Table 4. Study and population characteristics of the 15 randomized controlled trials comparing alternative surgical interventions for management of lumbar radiculopathy (EQ1) (continued)






80% power, = 5%)a;

Other outcomes

Franke (2009)34 Germany; Some concerns

Age 44 (11.7) Women 40 (40%) Duration of symptoms: NR

Microscopically assisted percutaneous nucleotomy N randomized: 52; N analyzed: 52 (100%); N crossovers: 0 (0%)


Operation duration (15 minutes);

VAS (sum of leg and back pain)

N (%) with motor deficits

N (%) with sensory deficits


Duration of postoperative inability to work

Haines (2002)40 United States; High

Age SG1: 42.2 (12.0); SG2: 35.4 (10.1) Women SG1: 10 (47.6%); SG2: 5 (38.4%) Duration of symptoms: NR

Automated percutaneous discectomy, endoscopic percutaneous discectomy N randomized: 21; N analyzed: 17 (81.0%) at 26w; N crossovers: 0 (0%)

Discectomy N randomized: 13; N analyzed: 10 (76.9%) at 26w; N crossovers: 1 (7.69%)

Difference in success rates (15%);



N (%) with success

N (%) with success based on MacNab criteria

Henriksen (1996)33 Denmark; Some concerns

Age Median (IQR) SG1: 39.7 (30 to 46); SG2: 42.8 (36 to 48) Women; SG1: 15 (38.5%); SG2: 14 (35%) Duration of symptoms: NR


Discectomy N randomized: 40; N analyzed: 40 (100%); N crossovers: 0 (0%)

Hospital stay duration (one day);

VAS 100 leg pain

VAS 100 back pain

Hermantin (1999)41 United States; Some concerns

Age SG1: Mean 40 (range 18-67); SG2: Mean 39 (range 15-66) Women’s SG1: 13 (43.4%); SG2: 8 (26.7%) Duration of symptoms: minimum duration of nonoperative treatment prior to randomization was 14w in both groups.

Video-assisted arthroscopic microdiscectomy N randomized: 30; N analyzed 30 (100%); N crossovers 0 (0%)

Discectomy, with laminotomy N randomized: 30; N analyzed: 30 (100%); N crossovers: 0 (0%)

NR (NR);

VAS 10 leg pain

N (%) with postoperative reflex abnormalities


N (%) with motor weakness

Duration of postoperative disability in time lost from work or until able to resume normal activity

N (%) self-reported satisfied with operative result

N (%) with satisfactory outcomes

Huang (2005)23 Taiwan; Some concerns

Age SG1: 39.2 (10.8); SG2: 39.8 (11.0) Women SG1: 4 (40%); SG2: 3 (25%) Duration of symptoms: NR

Microendoscopic discectomy N randomized: 10; N analyzed: 10 (100%); N crossovers: 0 (0%)


NR (NR);

N (%) with excellent/good outcome based on MacNab criteria

(continued)









80% power, = 5%)a;

Other outcomes

Mayer (1993)32 Germany; High

Age SG1: 39.8 (10.4); SG2: 42.7 (10) Women SG1: 8 (40%); SG2: 6 (30%) Duration of symptoms, mean (SD) SG1: 27.6w (NR); SG2: 29.2w (NR)

Percutaneous endoscopic discectomy N randomized: 20; N analyzed: 20 (100%); N crossovers: 3 (15%)


NR (NR);

N (%) with low back pain

N (%) with sciatica

N (%) with sensory deficit

N (%) with motor deficit

Duration of postoperative disability

N (%) returning to work

Clinical score

N (%) with specified clinical score

N (%) with self-reported specified success of surgery

Ruetten (2008)29 Germany; High

Age: 43 (range 20 to 68) Women: 116 (58%) Duration of symptoms, mean (range): 11.71w (0.14 to 68)

Endoscopic (interlaminar or transforaminal) discectomy N randomized: 100; N analyzed: 91 (91%); N crossovers: 0 (0%)


NR (NR);

VAS 100 leg pain

VAS 100 back pain

NASS pain score

NASS neurology score


N (%) with no leg pain

N (%) with leg pain occasionally or pain was greatly reduced

N (%) with no improvements in leg pain

N (%) with progredient back pain

Duration of postoperative work disability

(continued)









80% power, = 5%)a;

Other outcomes

Ryang (2008)28 Gempt (2013)98 Germany; High

Age SG1: 38.2 (9.3); SG2: 39.1 (11.3) Women; SG1: 17 (56.7%); SG2: 11 (36.7%) Duration of symptoms: NR

Trocar microdiscectomy N randomized: 30; N analyzed: unclear; N crossovers: 0 (0%)

Microdiscectomy N randomized: 30; N analyzed: unclear; N crossovers: 0 (0%)

NR (NR);

VAS 10 pain

VAS 10 for improvement

SF-36 bodily pain


SF-36 physical component summary and mental component summary



N (%) with radicular pain


N (%) with motor deficits

Sasaoka (2006)24 Japan; High

Age 42.4 (range 20 to 72) SG1: 36.5 (range 25 to 60) SG2: 37.7 (range 20 to 58) Women 14 (42.4%) SG1: 9 (60.0%) SG2: 3 (27.3%) Duration of symptoms: NR

Microendoscopic discectomy N randomized: 15; N analyzed: unclear; N crossovers: 0 (0%)

Microdiscectomy N randomized: 11; N analyzed: unclear; N crossovers: 0 (0%)

NR (NR);

Japanese Orthopaedic Association Score

N (%) with residual low back pain or lumbar discomfort

Teli (2010)27 Italy; Some concerns

Age: 39.3 (range 27 to 61) Women: 73 (34.4%) Duration of pain, mean (SD): SG1: 11w (5) SG2: 12w (6)

Microendoscopic discectomy N randomized: NR; N analyzed: 70; N crossovers: 0 (0%)

Microdiscectomy N randomized: NR; N analyzed: 72; N crossovers: 0 (0%)

VAS10 leg pain and VAS 10 back pain (1.5 points);



(continued)









80% power, = 5%)a;

Other outcomes

Thome (2005)26 Barth (2008)99 Germany; Some concerns

Age SG1:42 (9); SG2: 40 (10) Women SG1: 18 (42.9%); SG2: 19 (45.2%) Duration of symptoms, mean (SD) SG1: 11w (12) SG2: 8w (10)

Sequestrectomy N randomized: 42; N analyzed: 42 (100%); N crossovers: 0 (0%)


NR (NR);

VAS10 leg pain

VAS10 back pain

SF-36 bodily pain




N (%) with total Prolo score >=7

N (%) with specified total Prolo score

N (%) with improvement in sensory deficit

N (%) with improvement in motor deficit

N (%) with specified change in sensory index

N (%) with specified change in motor deficit

N (%) with specified categories or impairment at work

N (%) with specified patient satisfaction index scores

Microdiscectomy compared with discectomy

Henriksen (1996)33 Denmark; Some concerns

Age Median (IQR) SG1: 39.7 (30 to 46); SG2: 42.8 (36 to 48) Women; SG1: 15 (38.5%); SG2: 14 (35%) Duration of symptoms: NR



Hospital stay duration (one day);

VAS 100 leg pain

VAS 100 back pain

Teli (2010)27 Italy; Some concerns

Age: 39.3 (range 27 to 61) Women: 73 (34.4%) Duration of pain, mean (SD): SG2: 12w (6) SG3: 11w (5)

Microdiscectomy N randomized: NR; N analyzed: 72; N crossovers: 0 (0%)

Discectomy N randomized: NR; N analyzed: 70; N crossovers: 0 (0%)

VAS10 leg pain and VAS 10 back pain (1.5 points);



(continued)









80% power, = 5%)a;

Other outcomes

Tullberg (1993)25 Sweden; Some concerns

Age SG1: 40 (range 17 to 59); SG2: 38 (range 18 to 64) Women SG1: 12 (40%); SG2: 9 (30%) N (%) with specified duration of symptoms: SG1: <4w:1 (NR) SG2: <4w: 0 (NR)



NR (NR);

VAS 10 leg pain

VAS 10 back pain

Postoperative sick leave

N (%) of patients out of work

N (%) with specified option on recovery


Abbreviations: IQR = interquartile range; N = number; NASS = North American Spine Society; NR = not reported; SD = standard deviation; SF-36 = short-form 36 survey; SG =

surgical group; VAS = visual analogue scale; w = week(s); y = years(s).



The interventions and comparators evaluated in studies are summarized in Table 5. The

comparator interventions used in the seven RCTs evaluating the efficacy of surgery compared to

nonsurgical management varied. One RCT compared the efficacy of microdiscectomy with

spinal manipulation (formalized protocol, mean number of treatment sessions was 21).22 One

RCT compared the efficacy of microdiscectomy with a specific physiotherapy protocol.31 One

RCT compared percutaneous disc decompression with coblation technology to epidural steroid

injection.39 The other four RCTs compared discectomy, microdiscectomy, or percutaneous

discectomy to ‘conservative management’ as needed and directed by treating clinicians.5,21,30,35

Conservative management may have included analgesic or anti-inflammatory medication, bed

rest, physical therapy, home exercise instruction, and education or counseling about the natural

course of the disease.

Table 5. Surgical and comparator interventions used among 22 included studies for EQ1

Surgical Interventiona Comparator Interventionb

Eff

icac

y R

CT

s

(k=

7)

Microdiscectomy Spinal manipulation22 Physiotherapy31

Percutaneous disc decompression with coblation technology39

Epidural steroid injection

Percutaneous disc decompression35 Discectomy5,30 Discectomy/microdiscectomy21

Conservative management

Co

mp

arat

ive

effe

ctiv

enes

s

RC

Ts

(k=

15)

Tubular/trocar discectomy28,38 Automated percutaneous lumbar discectomy36 Percutaneous endoscopic discectomy32 Endoscopic interlaminar or transforaminal discectomy29 Microendoscopic discectomy24,27 Sequestrectomy26 Percutaneous laser disc decompression37 Microscopically assisted percutaneous nucleotomy34

Microdiscectomy

Automated percutaneous discectomy/endoscopic percutaneous discectomy40 Video-assisted arthroscopic microdiscectomy41 cMicroendoscopic discectomy23,27

Discectomy

cMicrodiscectomy25,27,33 Discectomy a In the Appendix C Evidence Tables, these interventions are considered the surgical group and are denoted as SG1.

b In the Appendix C Evidence Tables, these interventions are considered the comparator groups; surgical comparators are denoted

as SG2 or SG3. Nonsurgical comparator groups are denoted as NS1.

c This study was a three-arm RCT that allocated participants to microendoscopic discectomy, microdiscectomy, and standard

discectomy; thus, it contributes to three comparisons of interest for this HTA.

Abbreviations: RCT = randomized controlled trial.

The surgical interventions in the 15 RCTs evaluating the comparative effectiveness of surgical

interventions also varied. Ten RCTs evaluated various minimally invasive and/or endoscopic

surgical procedures to microdiscectomy.26-29,32,34,36-38 Four RCTs evaluated various minimally

invasive surgical procedures to standard discectomy.23,27,40,41 Two RCTs compared

microdiscectomy with discectomy.25,33

Across the 22 included RCTs, studies reported outcomes at various time points spanning from

immediately postoperative to up to 10 years postoperative; no single efficacy measure was used

consistently across all included studies. The most common measures of pain included patient-



reported visual analog scales (VAS) of leg pain, back pain, or both; and the bodily pain subscale

of the Short Form 36 (SF-36). Measures of functioning also varied across studies and most

commonly included validated patient-reported instruments including the physical functioning

subscale of the SF-36, the Roland-Morris Disability Questionnaire, and the Oswestry disability

index (ODI). Some studies also used the Prolo Scale, a measure based on an observer’s

assessment of the patient and includes a functional subscale and an economic subscale. A few

studies reported neurologic symptoms and overall health-related quality of life (QOL). Outcomes

related to “return to work” were variably defined and reported by just over half of the 22

included studies. Other outcomes reported included satisfaction with symptoms and degree of

recovery, both measured using Likert scales.

We rated one RCT as low risk of bias,43 10 RCTs as some concerns for bias,22,23,25-27,33,34,36,37,41

and 10 RCTs as high risk of bias.5,21,24,28-32,35,40 We rated one study as having some concerns for

bias for outcomes reported at 6 weeks but high risk of bias for outcomes reported at 12 weeks or

later because of high attrition at later follow-up time points (30% or greater in both study

groups). All but one study38 did not blind participants to treatment allocation, and since nearly all

studies relied on patient-reported outcomes, most studies had at least some concerns for bias

since knowledge of the assigned treatment may impact such outcomes. Studies rated as high risk

of bias generally used inadequate randomization and allocation concealment (e.g., use of even

/odd 35) or had moderate to extensive levels of crossover between treatment arms. For example,

in the Weinstein et al. RCT (Spine Patient Outcomes Research Trial [SPORT]), 46.1% of

participants allocated to surgery did not receive surgery by 26 weeks followup, and 36.3% of

participants allocated to conservative management received surgery.21

3.2.1.2 Findings

This section is organized by outcome: pain; neurological symptoms; quality of life;

functioning/disability (physical, social, or psychological), return to work, and other outcomes

(e.g., satisfaction with treatment outcome, recovery). Within each outcome domain, we

synthesized the seven RCTs comparing surgery to nonsurgical interventions (i.e., efficacy RCTs)

separate from the 15 RCTs comparing alternative surgical interventions (i.e., comparative

effectiveness). For consistency when reporting outcomes by group, we reported the finding from

the surgical intervention group first, followed by the nonsurgical comparator group for the

efficacy RCTs or the surgical comparator groups for the comparative effectiveness RCTs. All

results presented and synthesized are the intent-to-treat analyses for which participants are

analyzed in the groups to which they were allocated. Studies may have also reported “as treated”

and “per protocol” analyses; this is discussed further in the Discussion. We note that some

studies did not report the actual outcome measure values, between-group differences, or

statistical tests of significant difference; we use “Not Reported” (“NR”) to indicate when this

occurred. To further assist readers with interpretation of reported efficacy outcomes, Table 6

describes the outcomes reported by included studies, including how the outcome is assessed, the

range of possible scores, the directionality of the score, the minimally important clinical

difference reported in the literature, and the required sample sizes that we calculated to detect

various between-group differences with 80% power at an alpha level of 0.05.



Table 6. Summary of efficacy outcomes reported by included studies, including score range, minimally important clinical difference, and required sample size to detect various between-group differences

Instrument Administration Score Range Interpretation of Between-Group Treatment Effecta

Minimally Important Difference from Literatureb

Between-Group Difference (Δ) and Corresponding Sample Size Requirements (N)c

VAS 100 mmd Leg or Back Pain

Patient reported 0 to 100 Higher scores represent more severe symptoms

Negative absolute mean difference favors intervention group

Between 7 to 11 points9,46,100

Δ 15: N=70 Δ 11: N=128 Δ 7: N=314 (assuming SD of 22)

SF-36 Bodily Pain subscale

Patient reported, 2 items from the SF-36 instrument

0 to 100 (norm-based: mean 50, SD (10)) Higher scores represent less severe symptoms

Positive absolute mean difference favors intervention group

MID for within-group improvement ranges from 14 to 38;100,101 MID is 3 for between-group differences 46,102

Δ 10: N=128 Δ 7: N=260 Δ 3: N=1,398 (assuming SD of 20) SF-36 Physical

Functioning subscale Patient reported, 10 items from the SF-36 instrument

SF-36 Physical Component Summary

Patient reported, scores multiplied by subscale factor score coefficients and summed over all 8 subscales

MID for between-group differences: 2 points for PCS 3 points for MCS.46,102

Δ 5: N=84 Δ 3: N=226 Δ 2: N=506 (assuming SD of 8)

SF-36 Mental Component Summary

Roland Morris Disability Questionnaire

Patient reported, 24 items with yes/no responses

1 to 24 Higher scores represent worse functional status

Negative mean difference favors intervention group

Between 2 to 5 points9,46

Δ 8: N=16 Δ 5: N=34 Δ 4: N=52 Δ 2: N=200 (assuming SD of 5)

Oswestry Disability Index

Patient reported, 10 items with 6-point Likert Scale

0 to 100 Higher scores represent worse functional status


Between 30% to 50% relative change, or absolute change of 5 to 17 points9,46,103,104

Δ 17: N=30 Δ 15: N=38 Δ 10: N=84 Δ 5: N=324 (assuming SD of 16)

Sciatica Index-Bothersomeness subscale and Frequency subscale

Patient reported 6-point Likert scale 0 (not bothersome) to 6 (extremely bothersome)

0 to 24 Higher scores represent more severe symptoms


None established, a 10% relative difference (difference 2.4) is probably reasonable.

Δ 4: 42 Δ 2.4: 114 Δ 1: 638 (assuming SD 4.5)

(continued)



Table 6. Summary of efficacy outcomes reported by included studies, including score range, minimally important clinical difference, and required sample size to detect various between-group differences (continued)

Instrument Administration Score Range Interpretation of Between-Group Treatment Effecta

Minimally Important Difference from Literatureb

Between-Group Difference (Δ) and Corresponding Sample Size Requirements (N)c

Prolo Score-Economic and Functional subscales

Clinician assessor using 4-point Likert scale 0 (worse) to 4 (best)

Higher score represents increased ability to work or better functioning

Positive mean difference favors intervention group

None established, a 10% relative difference (0.4) is probably reasonable.

Δ 0.8: N=52 Δ 0.4: N=200 Δ 0.2: N=788 (assuming SD of 1.0)

Δ = between-group difference in means

a Treatment effect is difference in mean scores at follow-up time point, with or without adjustment for baseline scores.

b From the broader musculoskeletal pain literature; we identified no studies establishing between-group MIDs specific to lumbar radiculopathy.

c We calculated the sample size requirements based on 80% power, two-tailed test, alpha level = 0.05 using STATA version 14.0 with the standard deviation indicated in the table

cell. The standard deviation used represents the median standard deviation of the measure at baseline for the studies included in this HTA. The sample size in italics is the sample

size we used to assess the imprecision domain for strength of evidence ratings; bodies of evidence generally not meeting the lower end of the MID threshold were downgraded one

level (not serious to serious); bodies of evidence not meeting the upper end of the MID threshold were downgraded by two levels (not serious to very serious).

d Also applicable to VAS 10 cm, between-group differences in VAS 100 mm can be divided by 10 to be applicable to VAS 10 cm.

Abbreviations: MID = minimally important difference; N = number; SF-36 = short-form 36 survey; SD = standard deviation; VAS = visual analogue scale.



Pain


Seven RCTs reported at least one pain outcome. Five were rated as high risk of bias,5,21,30,31,35

one was rated as some concerns for bias,22 and one was rated as high risk of bias for outcomes

later than 12 weeks and some concerns for bias for outcomes less than 12 weeks.39 Pain

outcomes reported included the VAS 100 mm or 10 cm for leg pain, the VAS 100 mm or 10 cm

for back pain, the SF-36 bodily pain subscale, the Sciatica index, the McGill Pain Questionnaire,

the Abderdeen back pain scale. A few studies also reported the frequency and proportion of

participants reporting reduced pain, no pain, or relief from pain. We were unable to conduct

quantitative synthesis for pain outcomes within this comparison because of outcome

heterogeneity and because some studies did not report measures of variance needed to conduct a

meta-analysis. Table 7 summarizes the findings and strength of evidence related to pain

outcomes for this comparison. A detailed description of findings follows this table.

Table 7. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for pain in persons with symptomatic lumbar radiculopathy (EQ1)

Certainty Assessment

Summary of Findings CERTAINTY/ Direction of Effect

№ of Studies

Risk of Bias

Inconsistency Indirectness Imprecision

Pain(leg)-VAS 100 mm (short- and medium-term) (followup: range 6 weeks to 26 weeks) (MID 7 to 11 points)

3 RCTs Very seriousa

Not serious Not serious Not serious Pain improved in both treatment groups. Between-group differences ranged from -6 to -26 points, all favoring surgery. Peul et al.30 (N=283) 8w AMD -17.7 (95% CI -23.1 to -12.3), 26w AMD -6.1 (95% CI -10.0 to -2.2); Osterman et al.31 (N=56) (calculated 6w AMD -17, calculated 26w AMD -13); Gerszten et al.39 (N=90) calculated 6w AMD -21 (p=0.002), calculated 12w AMD -23 (p=0.0001), calculated 26w AMD -26 (P=0.0008).

⨁⨁◯◯

LOW Favors surgery

Pain (leg)-VAS 100 mm (long-term) (followup: range 52 weeks to 5 years) (MID 7 to 11 points)

2 RCTs Very seriousb

Not serious Not serious Seriousc Improvements in pain persisted over time in both treatment groups. However, between-group differences were minimal by 52 w and beyond. Peul et al.30 (N=283) 52w AMD 0 (95% CI -4.0 to 4.0); 2y AMD 2 (95% CI -2.0 to 6.0), similar at 5y. Osterman et al.31 (N=56) calculated 52w AMD -7, calculated 2y AMD -13, repeated measures AMD from 0 to 2 years -9.0 (95%CI -20 to 1).

⨁◯◯◯

VERY LOW No difference

Pain (back)-VAS 100 mm (short- and medium term) (followup: range 6 weeks to 26 weeks) (MID 7 to 11 points)


Not serious Not serious Not serious Pain improved in both treatment groups. Between-group differences ranged from -7 to -24 points, most favoring surgery. Peul et al.30 (N=283) 8w AMD -11.3 (95% CI, -17.4 to -5.6), 26w AMD -2.3 (95% CI, -8.2 to 3.6). Osterman et al.31 (N=56) calculated 6w, 12w, and 26w AMD -13; Gerszten et al.39 (N=90) calculated 6w AMD -19 (p=0.0005), calculated 12w AMD -24 (P=0.0001); calculated 26w AMD -21 (P=0.002).

⨁⨁◯◯

LOW Favors surgery

(continued)



Table 7. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for pain in persons with symptomatic lumbar radiculopathy (EQ1) (continued)



№ of Studies

Risk of Bias


Pain (back)-VAS 100 mm (long-term) (followup: range 52 weeks to 5 years) (MID 7 to 11 points)


Not serious Not serious Seriousc Improvements in pain persisted over time in both treatment groups. However, between-group differences were minimal by 52 weeks and beyond. Peul et al.30 (N=283) 52w AMD -2.3 (95% CI -8.2 to 3.6), 2y AMD -1.4 (95% CI, -6.3 to 4.5), 5y AMD 3.1 (95% CI, -4.2 to 10.3). Osterman et al.31 (N=56) calculated 52w AMD -4, calculated 2y AMD -16, repeated measures AMD from 0 to 2y -7.0 (95%CI -17 to 3).

⨁◯◯◯


Pain- SF-36 bodily pain subscale (short- and medium-term) (followup: range 8 weeks to 26 weeks) (MID 3 points)

4 RCTs Very seriousd

Not serious Not serious Seriouse Pain improved in both treatment groups; between-group differences favored surgery at some time points in some studies. Weinstein et al.[SPORT]21 (N=501) 12w AMD 2.9 (95% CI, -2.2 to 8.0). Peul et al.30 (N=283) 8w AMD 8.4 (95% CI 3.2 to 13.5), 26w AMD 3.3 (95% CI, -1.8 to 8.4). Gerszten et al.39 (N=90) significant between-group difference favoring plasma disc decompression at 26 w (actual values NR, P=0.0039). McMorland et al.22 (N=40) no difference in repeated measures AMD 0 to 12w (Actual Values NR, P=0.341).

⨁◯◯◯

VERY LOW Favors surgery

Pain- SF-36 bodily pain subscale (long-term) (followup: range 52 weeks to 8 years) (MID 3 points)


Not serious Not serious Seriousf Improvements in pain persisted over time in both treatment groups; no between-group differences between observed in either study. Peul et al.30 (N=283) 52w AMD 2.7 (95% CI, -2.6 to 7.9), 2y AMD -2.3 (95% CI, -7.3 to 2.7). Weinstein et al.[SPORT]21 (N=501) 52w AMD 2.8 (95% CI, -2.3 to 7.8), 2y AMD 3.2 (95% CI -2.0 to 8.4), 4y AMD 4.5 (95% CI, -1.2 to 10.3), 8y AMD 0.7 (95% CI, -5.2 to 6.6); repeated measures AMD at 0 to 2y, 4y, and 8y were also nonsignificant.

⨁◯◯◯


Pain-Sciatica index (short- and medium-term) (followup: range 8 weeks to 26 weeks) (MID 2.4 points)

3 RCTs Very seriousg

Not serious Not serious Not serious Scores improved in both treatment groups; between-group differences favoring surgery in all studies at most follow-up time points. Bothersomeness subscale: Osterman et al.31 (N=56) 8w AMD -3.6 (95% CI, -4.9 to -2.3), 26w AMD -1.2 (95% CI, -1.3 to -0.1); Weinstein et al.[SPORT]21 (N=501) 12w AMD -2.1 (95% CI, -3.4 to -0.9). Frequency subscale: Peul et al.30 (N=283) 8w AMD -4.0 (95% CI, -5.3 to -2.7), 26w AMD -0.5 (95% CI, -1.8 to 0.8).

⨁⨁◯◯

LOW Favors surgery

(continued)



Table 7. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for pain in persons with symptomatic lumbar radiculopathy (EQ1) (continued)

Certainty Assessment Summary of Findings CERTAINTY/ Direction of Effect

Pain- Sciatica index (long-term) (followup: range 52 weeks to 8 years) (MID 2.4 points)

3 RCTs Very seriousg

Not serious Not serious Not serious Improvements in scores persisted over time in both treatment groups; between-group differences varied by study. Bothersomeness subscale: Osterman et al.31 (N=56) 52w AMD -0.4 (95% CI, -1.5 to 0.7); Weinstein et al.[SPORT]21 52w AMD -1.6 (95% CI, -2.9 to -0.4), and significant differences at 2y, 4y, and 8y though AMDs all less than the MID. Frequency subscale: Peul et al.30 (N=283) 52w AMD -0.5 (95% CI, -1.8 to 0.8).

⨁⨁◯◯

LOW No differenceh

a Risk of bias was high in all trials for medium-term outcomes; risk of bias was some concerns in 1 trial for short-term outcomes.

Sources of bias: lack of participant and outcome assessor blinding and differential crossover.

b Risk of bias was high in both trials. Sources of bias: lack of participant and outcome assessor blinding and differential

crossover.

c Optimal information size criteria not met in smaller trial31 sample size can only detect between-group differences of more than

15 points.

d Risk of bias was rated as high in 3 trials, and as some concerns in 1 trial.22 Sources of bias: lack of participant and outcome

assessor blinding and extensive and differential crossover.

e Optimal information size criteria not met by some trials: a sample size of 1,398 is required to detect a difference of 3 points,

actual values and CI for AMDs not provided by some studies.

f Optimal information size criteria not met: a sample size of 1,398 is required to detect a difference of 3 points

g Risk of bias was high in all trials. Sources of bias: lack of participant and outcome assessor blinding and extensive and

differential crossover.

h Although statistically significant between-group differences were observed in one trial)21 at all follow-up time points; the

AMDs were all less than the MID for this measure; thus we concluded a low strength of evidence for no meaningful clinical

difference.

Abbreviations: AMD = absolute mean difference; CI = confidence interval; MID = minimally important between-group

difference; NR = not reported; RCT = randomized controlled trial; SPORT = Spine Patient Outcomes Research Trial; VAS =

Visual Analog Scale; SF-36 = short-form 36 survey; w = week(s); y = year(s)

VAS Leg Pain

Three RCTs reported leg pain outcomes using VAS 100 mm scales. A higher score on this scale

indicates worse pain, and a negative AMD favors the surgical intervention over the nonsurgical

comparator intervention. Peul et al.,30 Osterman et al.,31 and Gerstzen et al.,39 reported decreased

scores (i.e., improvement) for leg pain in participants allocated to both the surgical treatment and

nonsurgical comparator from baseline through short-term (6 and 8-week followup) and medium-

term (up to 26 weeks) followup. Between-group differences in VAS scores ranged from -6 to -26

points favoring surgery at short- and medium-term followup. Of the 2 studies reporting long-

term outcomes, improvements in pain persisted but between-group differences were not

significant. Specific study findings:

In Gerstzen et al., between-group differences at 6 weeks, 12 weeks, and 26 weeks were

significantly different, favoring plasma disc decompression compared with epidural steroid



injection (12-week mean (SD) decrease in VAS 100 mm leg pain: -46 (4) in surgical group,

-23 (5) in epidural steroid group, P=0.0001).39

Peul et al. reported differences between groups for short-term and medium-term followup

that favored microdiscectomy over conservative management (8-week AMD -17.7 (95%

CI, -23.1 to -12.3; 26-week AMD -6.1 (95% CI, -10.0 to -2.2).30 However, differences

between groups were minimal and not statistically different by 52 weeks, 2 years, and 5

years (2-year AMD: 2.0 [95% CI, -2.0 to 6.0]).

Osterman et al. also reported larger decreases among participants allocated to surgery

through short-term and medium-term followup but no tests of significance at these time

points were reported.31 The calculated AMD at 6 weeks, 12 weeks, and 26 weeks was -17, -

11, and -13, respectively. Differences between groups were minimal at 52 weeks

(calculated AMD -7), but larger decreases were observed again at 2 years among

participants allocated to surgery (calculated AMD -13). The study reported a repeated

measures AMD between groups from 0 to 2 years of -9 (95% CI, -20 to 1).

VAS Back Pain

The same three RCTs also reported back pain outcomes using VAS 100 mm scales.30,31,39 In all

three RCTs, VAS back pain scores were lower than VAS leg pain scores at baseline, consistent

with a study population selected for radicular pain. Back pain scores decreased among

participants allocated to both the surgery and the nonsurgical comparator groups in all RCTs.

The between-group differences in VAS back pain scores in the short and medium term ranged

from -7 to -24. Of the 2 studies reporting long-term outcomes, improvements in pain persisted

but between-group differences were not significant. Specific study findings:

Peul et al. reported an AMD between microdiscectomy and conservative management of -

11.3 (95% CI, -17.4 to -5.6) at 8 weeks, -2.3 (95% CI, -8.2 to 3.6) at 26 and 52 weeks, and

-1.4 (95% CI, -6.3 to 4.5) at 2 years.30

Osterman et al. reported larger decreases in participants allocated to microdiscectomy

compared to physiotherapy at 6 weeks, 12 weeks, and 26 weeks (calculated AMD -13 at all

time points), but no tests of significant differences were reported.31 The difference between

groups was minimal at 52 weeks (calculated AMD -4), but larger decreases among

participants allocated to surgery were observed again at 2 years (calculated AMD -16). The

repeated measures AMD from 0 to 2 years was not statistically significant (AMD -7 (95%

CI -17 to 3).

In Gerstzen et al., VAS back pain scores remained similar or increased from baseline

(range from decrease of -0.4 to increase of 7 at various time points) in participants

allocated to epidural steroid injection while scores decreased in participants allocated to

plasma disc decompression (range of decrease -18 to -21 at different time points).39 This

resulted in between-group differences that were significant at 6 weeks (P=0.002), 12 weeks

(P=0.001), and 26 weeks (P=0.0008).



VAS Pain

One RCT also reported pain outcomes using VAS, but it was not specific to leg or back pain.

Erginousakis et al. used a 10-cm VAS outcome to assess the effectiveness of percutaneous disc

decompression to conservative management.35 Similar to the other three RCTs, VAS scores

decreased from baseline to followup in participants allocated to both the surgery (baseline mean

[SD] 7.4 [1.4]) and nonsurgical intervention (baseline mean [SD] 6.9 [1.9]. At 12 weeks

followup, the mean (SD) among participants allocated to percutaneous disc decompression was

3.0 (2.4) and was 0.9 (2.0) among participants allocated to conservative management; this

difference was not statistically different (P> 0.005 [sic]). Pain scores increased among

participants allocated to conservative management and decreased among participants allocated to

surgery at 52 weeks and 2 years resulting in significant between-group differences favoring

surgery (P=0.005 and P=0.004, respectively). We did not use this study in the strength of

evidence ratings for the VAS pain outcome because of the lack of specificity for leg versus back

pain.

SF-36 Bodily Pain

Four RCTs reported pain outcomes using the bodily pain subscale of the SF-36; on this scale

higher scores represent less pain and a positive between-group AMD favors the surgical

intervention relative to the nonsurgical intervention. The studies all reported increases in scores

from baseline to short- and medium-term followup among participants allocated to both the

surgical intervention and the nonsurgical comparator; the range of increase was 14 to 42 points.

However, between-group differences at followup varied by study. Specific study findings:

Gerszten et al. reported larger improvements in participants allocated to plasma disc

decompression compared to participants allocated to epidural steroid injection at 26 weeks

(actual values NR, P=0.0039).39

Weinstein et al. [SPORT] reported nonsignificant difference at all follow-up time points

between 12 weeks and 8 years (12-week AMD 2.9 [95% CI, -2.2 to 8.0]; 2-year AMD 3.2

[95% CI, -2.0 to 8.4]) and also reported nonsignificant repeated measures differences at 2

years (AMD NR, P=0.74), 4 years (AMD NR, P=0.15), and 8 years (AMD NR,

P=0.22).21,44,45

McMorland et al. also reported a nonsignificant repeated measures difference from 0 to 12

weeks (AMD NR, P=0.341).22

Peul et al. reported a significant difference at 8 weeks favoring surgery (AMD 8.4 [95%

CI, 3.2 to 13.5]). However, the outcomes reported at 26 weeks (AMD 3.3 [95% CI, -1.8 to

8.4]), 52 weeks (AMD 2.7 [95% CI, -2.6 to 7.9]), and 2 years (AMD -2.3 [95% CI, -7.3 to

2.7]) showed no significant between-group differences.30,95

Sciatica Index

Three RCTs reported outcomes measured with the Sciatica index.21,30,31 With this measure,

higher scores represent worse symptoms and negative between-group difference favor the

surgical intervention. Similar to other pain outcomes already reported, scores improved over time

in both the surgical and nonsurgical intervention groups. Between-group differences favored



surgical interventions compared with nonsurgical interventions at short- and medium-term

followup. Specific study findings:

Osterman et al. reported an AMD of -3.6 (95% CI, -4.9 to -2.3) at 8 weeks and -1.2 (95%

CI, -1.3 to -0.1) at 26 weeks for the Bothersomeness subscale of the Sciatica index, which

favored surgery compared with physiotherapy.31 By 52 weeks, the difference between

groups was not significant (AMD -0.4 [95% CI, -1.5 to 0.7]).

Weinstein et al. [SPORT] reported an AMD of -2.1 (95% CI, -3.4 to -0.9) at 12 weeks

favoring surgery compared to conservative management for the same subscale.21

Significant differences favoring surgery persisted at all additional follow-up time points (52

weeks, 2 years, 4 years, and 8 years) and in repeated measures analyses from 0 to 2 years,

and from 0 to 8 years.21,44,45

Peul et al. reported outcomes for the Frequency subscale of this index.30 The AMD

between groups favored microdiscectomy compared with conservative management at 8

weeks (AMD -4.0 [95% CI, -5.3 to -2.7]) and 26 weeks (AMD -1.8 [95% CI, -1.9 to -0.7]);

this benefit did not persist at 52 weeks (AMD -0.5 (95% CI, -1.8 to 0.8]) and was not

reported at 2 years or 5 years of followup.

Other Pain Measures

Three RCTs reported other measures related to pain.5,22,35 These measures confirmed similar

findings to other measures of pain previously reported. We did not use these studies in our

strength of evidence ratings for VAS pain outcome because they were only used in one study

each. Specific study findings:

McMorland et al. reported pain using three different subscales of the McGill Pain

Questionnaire and the Abderdeen Back Pain Scale at 12 weeks followup.22 Outcomes

measured using the three McGill Pain subscales found that pain decreased significantly

over time in participants allocated to both microdiscectomy or spinal manipulation. Though

a numerically larger decrease was observed in the surgical group, the differences in the

repeated measures AMD were not statistically significant. In contrast, use of the Abderdeen

Back Pain Scale found that participants allocated to microdiscectomy had a significantly

larger decrease in pain compared to spinal manipulation (AMD NR, P=0.034).

Weber et al. reported the frequency and proportion of participants within three categories

of radiating pain at 4 years and 8 years (no pain, some pain, considerable pain).5 The

proportion of participants with no pain was 63.2% among those allocated to discectomy

and 57.6% among those allocated to conservative management at 4 years followup

(calculated P=0.86). At 10 years, these proportions were 84.3% and 78.8%, respectively

(calculated P=0.41).

Erginousakis et al. reported the proportion of participants within four categories of pain

reduction at 2 years (100% pain relief, 50% pain relief, 0% pain relief, and aggravation of

pain).35 The proportion of participants that reported 100% pain relief was 55% among those

allocated to percutaneous disc decompression and 19% among those allocated to

conservative therapy (calculated P=0.008).




Ten RCTs comparing minimally invasive surgical interventions (tubular/trocar discectomy,28,38

percutaneous endoscopic discectomy,32 endoscopic interlaminar or transforaminal discectomy,29

microendoscopic discectomy,24,27 sequestrectomy,26 percutaneous laser disc decompression,37

microscopically assisted percutaneous nucleotomy,34 and video-assisted microdiscectomy41) to

either microdiscectomy or discectomy reported at least one pain outcome. Four were rated as

high risk of bias,24,28,29,32 five were rated as having some concerns for bias,26,27,34,37,41 and one

was rated as low risk for bias.38 Pain outcomes reported include the VAS 100 mm or 10 cm for

leg pain or back pain, the SF-36 bodily pain subscale, the Sciatica index, and the frequency and

proportion of participants reporting reduced pain, no pain, or relief from pain. Table 8

summarizes the findings and strength of evidence related to pain outcomes for this comparison.

A detailed description of findings follows this table.

Table 8. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for pain in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Pain (leg)-VAS 100 mm (short-term) (followup: range 4 weeks to 8 weeks) (MID 7 to 11 points)

2 RCTs Not serious

Not serious Not serious Seriousa Decrease in scores from baseline in both surgical groups. Between-group differences were not significant in either study. Arts et al.38 (N=328) 4w AMD 4.5 (95% CI, -0.3 to 9.3), 8w AMD 4.5 (95% CI, -0.4 to 9.3). Brouwer et al.37 (N=115) 4w AMD 7.4 (95% CI, -1.9 to 16.8), 8w AMD 5.7 (95% CI, -3.7 to 15.0).

⨁⨁⨁◯

MODERATE No difference

Pain (leg) VAS 100mm and 10 cm (medium-term) (followup: range 12 weeks to 26 weeks) (MID 7 to 11 points)

5 RCTs Seriousb Not serious Not serious Not serious Decrease in scores from baseline in both surgical groups. Between-group differences were not significant in any study.26-28,37,38,41 Pooled between-group mean difference in scores at followup 0.3 (95% CI, -2.2 to 2.9, 4 RCTs, 642 participants, I2=0%).c

⨁⨁⨁◯


Pain (leg) VAS 100mm and 10 cm (long-term) (followup: range 52 weeks to 2 years) (MID 7 to 11 points)

5 RCTs Seriousb Not serious Not serious Not serious Improvements in pain persist in both surgical groups in all studies; however, between-group differences not significant at any single long-term followup point.26,27,29,37,38 Pooled between-group mean difference in scores at 52w to 1.5 y 1.6 (95% CI, -1.5 to 4.6, 4 studies, 640 participants, I2=28.1%) and at 2y -0.1 (95% CI, -2.7 to 2.4, 4 RCTs, 619 participants, I2=0%).c Arts et al.38 (N=328) repeated measures: 0 to 52w AMD 4.2 (95% CI 0.9 to 7.5), 0 to 2y AMD 3.3 (95% CI 0.2 to 6.2). Brouwer et al.37 (N=115) repeated measures: 0 to 52w AMD 6.9 (95% CI, 1.3 to 12.6), 0 to 2y AMD 5.0 (95% CI, -0.2 to 10.2).

⨁⨁⨁◯


(continued)



Table 8. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for pain in persons with symptomatic lumbar radiculopathy (EQ1) (continued)



№ of Studies

Risk of Bias


Pain (back)-VAS 100mm (short-term) (followup: range 4 weeks to 8 weeks) (MID 7 to 11 points)

2 RCTs Not serious

Not serious Not serious Seriousa Decrease in scores from baseline in both surgical groups in both studies. Between-group differences were not significant in either study. Arts et al.38 (N=328) 4w AMD 3.1 (95% CI, -1.9 to 8.1), 8w AMD 3.8 (95% CI, -1.3 to 8.8). Brouwer et al.37 (N=115) 4w AMD -2.0 (95% CI, -11.3 to 7.2), 8w AMD 6.3 (95% CI, -2.9 to 15.5).

⨁⨁⨁◯


Pain (back) VAS 100mm and 10 cm (medium-term) (followup: range 12 weeks to 26 weeks) (MID 7 to 11 points)

5 RCTs Seriousb Seriousd Not serious Not serious Improvements in pain persist in both surgical groups in all studies.26,27,29,37,38 Between-group differences were significant in only 1 study: Brouwer et al.37(N=115) AMD 9.4 (95% CI, 0.1 to 18.6). Pooled between-group mean difference in scores at 12 to 26w was 1.3 (95% CI, -3.5 to 6.2, 4 RCTs, 642 participants, I2=61.7%).

⨁⨁◯◯

LOW No difference

Pain (back) VAS 100mm and 10 cm (long-term) (followup: range 52 weeks to 2 years) (MID 7 to 11 points)

5 RCTs Seriousb Seriousd Not serious Not serious Improvements in pain persist in both surgical groups in all studies.26-28,37,38 Between-group differences at any single followup were not significant. Pooled between-group mean difference in scores at 52w to 1.5y followup 1.5 (95% CI, -3.0 to 5.9, 4 RCTs 640 participants, I2=57.6%); pooled mean difference at 2y -0.8 (95% CI, -5.7, to 4.1, 4 RCTs, 619 participants, I2=59.0%. Repeated measures: Arts et al.38,43 (N=328) AMD 0 to 52w 3.5 (95% CI, 0.1 to 6.9), 0 to 2y and 3.0 (95% CI, -0.2 to 6.3). Brouwer et al.37,97 (N=115) AMD 0 to 52w 7.6 (95% CI, -1.7 to 16.9), AMD 0 to 2y -1.5 [95% CI, -11.0 to 8.0).

⨁⨁◯◯

LOW No difference

Pain-SF-36 bodily pain (short-term) (followup: range 4 weeks to 8 weeks) (MID 3 points)

2 RCTs Not serious

Seriouse Not serious Seriousf1 Increase in scores from baseline in both surgical groups; no significant between-group differences. Arts et al.38 (N=328) 4w AMD -1.6 (95% CI, -6.7 to 3.6), 8w AMD -5.1 (95% CI, -10.3 to 0.1); Brouwer et al.37 (N=115) 4w AMD 4.1 (95% CI, -4.8 to 12.9), 8w AMD 0.6 (95% CI, -9.1 to 9.3).

⨁⨁◯◯

LOW No difference

Pain-SF-36 bodily pain (medium-term) (followup: range 12 weeks to 26 weeks) (MID 3 points)

3 RCTs Seriousg1 Seriousd Not serious Serioush1 Improvements in pain persist in both surgical groups;26,37,38 1 of 3 studies reports a significant between-group difference (Brouwer et al.37 [N=115] 26w AMD -11.3 [95% CI, -20.1 to -2.4]). Pooled between-group mean difference at 12 to 26w -3.0 (95% CI, -12.8 to 6.8), 3 RCTs, 500 participants, I2=75.4%).

⨁◯◯◯


(continued)



Table 8. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for pain in persons with symptomatic lumbar radiculopathy (EQ1) (continued)



№ of Studies

Risk of Bias


Pain-SF-36 Bodily Pain (long-term) (followup: range 52 weeks to 2.8 years) (MID 3 points)

3 RCTs Seriousi1 Seriousd Not Serious Serioush1 Increase in scores from baseline to 52w, 2y, and 2.8y in both surgical groups in all studies: Arts et al.38 (N=328), Brouwer et al.37 (N=115), and Ryang et al.28 (N=60). Between-group differences ranged from -11 to 2.5 points; none of these between-group differences were statistically significant.

⨁◯◯◯


Sciatica index (short-, medium-, and long-term) (followup: range 4 weeks to 2 years) (MID 2.4 points)

2 RCTs Not Serious

Not serious Seriousj Not serious Decreases in scores from baseline to all follow-up time points in both surgical groups for both the Bothersomeness and Frequency subscales; no significant between-group differences at any single time point, or in repeated measures analyses in either study. Arts et al.38 (N=328) repeated measures AMD 0 to 52w 0.7 (95% CI, -0.1 to 1.5), 0 to 2y 0.5 (95% CI, -0.3 to 1.3) for Bothersomeness subscale; similar findings for Frequency subscale. See Appendix C, Table C-3 for findings related to Brouwer et al.37 (N=115).

⨁⨁⨁◯


a Optimal information size criterion not met: Smaller trial37 is unable to detect differences small than 11 points, the larger trial38 is

only able to detect differences of 7 points or more.

b 1 trial was rated as low risk of bias38, 3 were rated as some concerns for bias,26,27,41 and 1 was rated as high risk of bias28.

Sources of bias: lack of participant and outcome assessor blinding (in all but the 1 trial rated as low risk of bias) and inadequate

randomization and allocation concealment in the trial rated as high risk of bias.28

c 1 of the 5 trials did not provide measures of variance required for pooling effect estimate.

d I2 statistic indicates moderate heterogeneity; trials are inconsistent with respect to magnitude and direction of effect and overlap

in confidence intervals

e Confidence intervals are modestly wide and spanning thresholds that would favor both interventions, less overlap in confidence

intervals for 8w outcomes.

f Optimal information size criterion not met: a sample size of 1,398 required to detect a difference of 3 points; the smaller trial37

is unable to detect differences less than 11 points, and the larger trial38 is unable to detect differences of less than 7 points.

g 2 of 3 trials had some concerns for bias,26,37 the other trial was low risk for bias.38 Sources of bias: lack of participant and

outcome assessor blinding. Sources of bias: lack of participant and outcome assessor blinding.

h Optimal information size criterion not met: a sample size of 1,398 required to detect a difference of points, the pooled estimate

for medium-term outcomes is based on a sample size of 500. For long term outcomes, the largest trial38 would be unable to detect

a difference less than 7 points, and the two other studies would not be able to detect differences smaller than approximately 11

points.

i1 trial was low risk of bias,38 1 trial was some concerns for bias,37 and 1 trial was high risk of bias.28 Sources of bias: lack of

participant and outcome assessor blinding (all trials) and inadequate randomization and allocation concealment 28.

j One of the trials37 used an adapted approach to scoring this measure, it is unclear whether this adapted approach has the same


difference; N = number; NR = not reported; RCT = randomized controlled trial; SF-36 = short-form 36 survey; VAS = visual

analog scale; w = week(s); y = year(s)



VAS Leg and Back Pain

Five RCTs reported leg or back pain outcomes using the VAS 100 mm or 10 cm scale.26,27,29,37,38

A decrease in score represents improvements in pain, a negative AMD between groups favors

the minimally invasive surgical procedure. Two studies reported decreases in VAS 100 mm leg

and back pain scores from baseline to 4 weeks and 8 weeks followup among participants

allocated to minimally invasive surgery and among participants allocated to

microdiscectomy.37,38 Between-group differences in short-term outcomes were not significant in

either study for either leg or back pain. Specific study findings:

Arts et al. reported between group differences of 4.5 (95% CI, -3 to 9.3) at 4 weeks and 4.5

(95% CI, -0.4 to 9.3) at 8 weeks for VAS 100 mm leg pain scores.38 Similar findings for

VAS Back Pain.

Brouwer et al. reported between-group differences of 7.4 (95 % CI, -1.9 to 16.8) at 4 weeks

and 5.7 (95% CI, -3.7 to 15.0) at 8 weeks for VAS 100 mm leg pain scores.37

Nonsignificant between-group findings were also observed for VAS 100 mm back pain

scores.

All five RCTs reported medium-term outcomes using VAS 100 mm or 10 cm for leg and back

pain. With one exception, between-group differences were not significant in any study. The

pooled between-group difference in VAS 100 mm leg pain at 12 to 26 weeks was 0.3 (95% CI, -

2.2 to 2.9, 4 RCTs, 642 participants, I2=0%, Appendix F, Figure F-1). The pooled between-

group difference in VAS 100 mm back pain at 12 to 26 weeks was 1.3 (95% CI, -3.5 to 6.2, 4

RCTs, 642 participants, I2=61.7%, Appendix F, Figure F-2). Specific study findings:

Arts et al. and Brouwer et al. found a similar pattern for VAS 100 mm leg pain at 26

weeks followup: Arts et al.38 reported a between-group difference of 2.0 (95% CI, -2.9 to

6.8) and Brouwer et al.37 reported a between-group difference of 4.2 (95% CI, -5.2 to 13.6).

A similar finding was reported in Arts et al for VAS 100 mm back pain; however, Brouwer

et al., reported a statistically significant between-group difference of 9.4 (95%, CI, 0.1 to

18.6) at 26 weeks despite short-term and long-term outcomes that were not significant.

Ruetten et al. reported a decrease in VAS 100 mm leg pain score of 75 at baseline to 6 at

12 weeks and 9 at 26 weeks among participants allocated to endoscopic discectomy and a

decrease from 71 at baseline to 9 at 12 weeks and 7 at 26 weeks; between-group

differences were reported as nonsignificant.29

Teli et al. and Thome et al. reported leg pain outcomes using the VAS 10 cm scale. Teli et

al. reported a decrease from baseline to 26 weeks of 8 (SD 1) to 2 (SD 1) among those

allocated to microendoscopic discectomy, from 8 (SD 1) to 2 (SD 1) among those allocated

to microdiscectomy, and from 8 (SD 1) to 2 (SD1) among those allocated to standard

discectomy.27 At 12 to 26 weeks, Thome et al. reported a decrease from 5.9 (SD 2.6) to 0.7

(SD 1.7) among those allocated to sequestrectomy and from 6.7 (AD 2.3) to 1.3 (SD 2.5)

among those allocated to microdiscectomy. 26

All 5 RCTs also reported long-term outcomes at 52 weeks and 2 years with VAS 100 mm or 10

cm, and the reductions in leg and back pain observed in the short and medium term generally



persisted. Between-group differences were reported for both single points in followup and using

repeated measures in some studies. The pooled between-group difference in VAS 100 mm leg

pain at 52 weeks to 1.5 years was 1.6 (95% CI, -1.5 to 4.6, 4 RCTs, 640 participants, I2=28.1%,

Appendix F, Figure F-1). The pooled between-group difference in VAS 100 mm back pain at

52 weeks to 1.5 years was 1.5 (95% CI, -3.0 to 5.9, 4 RCTs, 640 participants, I2=57.6%

Appendix F, Figure F-2). Specific study findings:

For Arts et al. and Brouwer et al, between-group differences in VAS 100 mm leg pain at

any single followup at 52 weeks or 2 years were not significant; however, both studies

reported at least one significant between-group repeated measures difference. Arts reported

a repeated measures AMD of 4.2 (95% CI, 0.9 to 7.5) for 0 to 52 weeks and 3.3 (95% CI,

0.2 to 6.2) for 0 to 2 years; both estimates favoring microdiscectomy compared with

tubular discectomy.38,43 Brouwer et al. reported a repeated measure mean difference of 6.9

(95% CI, 1.3 to 12.6) for 0 to 52 weeks and 5.0 (95% CI, -0.2 to 10.2) for 0 to 2 years.37,97

A similar pattern was observed for VAS 100 mm back pain scores. Between-group

differences at any single followup at 52 weeks or 2 years were not significant but Arts et al.

reported a significant repeated measures mean difference of 3.5 (95% CI, 0.1 to 6.9) for 0

to 52 weeks favoring microdiscectomy compared with tubular discectomy but the repeated

measures AMD for 0 to 2 years indicated no between-group differences (AMD 3.0 [95%

CI, -0.2 to 6.3]).38,43 Brouwer et al. reported no significant repeated measures between-

group differences (AMD 4.6 [95% CI, -1.1 to 10.4] for 0 to 52 weeks, AMD 3.0 [95% CI, -

2.2 to 8.1] for 0 to 2 years).37,97

Ruetten et al.,29 Teli et al.,27 and Thome et al.26 all reported no significant between-group

differences at 52 weeks and 2 years for VAS 100 mm leg or back pain scores.

VAS Pain

Three RCTs reported on other variants of the VAS pain measure.28,34,41 These findings were

consistent with the previously reported VAS outcomes, but we did not use these studies in our

strength of evidence ratings for VAS pain outcome because of the lack of specificity for leg

versus back pain or because of unspecified follow-up times. Specific study findings:

Franke et al. reported a repeated measures sum of VAS 10 cm leg and back pain scores at

52 weeks.34 Although pain scores decreased over time in participants allocated to

microscopically-assisted percutaneous nucleotomy and among participants allocated to

microdiscectomy, significant between-group differences favoring the minimally invasive

surgical procedure (actual values NR, P=0.006) were only observed at one of the two

clinical centers that enrolled patients. Further, when the leg and pain scores were

considered in a post-hoc analysis, between-group differences were only seen for back pain.

Hermantin et al. reported outcomes using the VAS 10 cm scale, but did not specify whether

it was for leg pain or back pain, and the follow-up time period was also not specified.41

Pain decreased from 6.8 (SD NR) to 1.2 (SD NR) among participants allocated to video-

assisted arthroscopic microdiscectomy and from 6.6 (SD NR) to 1.9 (SD NR) among

participants allocated to discectomy with laminotomy. No statistical tests of between-group

differences were reported.



Ryang et al. reported VAS 10 cm pain scores (unspecified as to leg or pain) over long-term

followup. Scores decreased from baseline to both follow-up time points among participants

allocated to trocar discectomy and among patients allocated to microdiscectomy, but

between-group differences at 1.33 years (P=0.86) and 2.8 years (P value reported as NS)

suggested no differences.

SF-36 Bodily Pain

Four RCTs compared minimally invasive surgical interventions to microdiscectomy and reported

pain outcomes using the bodily pain subscale of the SF-36.26,28,37,38 For this outcome, a higher

score represents less severe pain and a positive between-group AMD favors the minimally

invasive surgical procedure. In all studies, pain scores improved from baseline to short-,

medium-, and long-term followup among participants allocated to both surgical groups.

Increases in scores ranged from 37 to 51 points over the various follow-up times. With one

exception, no between-group differences were observed at any follow-up time. The pooled mean

difference in SF-36 bodily pain scores at 12 to 26 weeks was -3.0 (95 % CI, -12.8 to 6.8, 3

RCTs, 500 participants, I2=75.4%, Appendix F, Figure F-3). Specific study findings:

Arts et al. and Brouwer et al. reported nonsignificant between-group differences at 4 weeks

and 8 weeks that ranged from -5.1 to 0.6 points.37,38 At 26 weeks, Arts et al. reported an

AMD of -4.9 (95% CI, -10.0 to 0.3) while Brouwer et al. reported a significant between-

group difference favoring microdiscectomy compared with percutaneous laser disc

decompression (AMD -11.3 [95% CI, -20.1 to -2.4]).37

Thome et al. also observed a nonsignificant, between-group difference (calculated AMD

3.0, P=0.14) at 12 to 26 weeks.26

Arts et al., Brouwer et al., Thome et al., and Ryang et al. reported nonsignificant between-

group differences ranging from -11 points to 2.5 points at 52 weeks to 2.8 years

followup.26,28,37,38

Sciatica Index

Two RCTs also reported outcomes measured with the Sciatica index.37,38 With this measure, a

higher score represents more severe symptoms and a negative between-group difference favors

the minimally invasive surgical approach. Similar to other pain outcomes already reported,

scores improved over time in both surgical groups. Between-group differences were

nonsignificant for both the Bothersomeness subscale and the Frequency subscale at all single

follow-up time points in both trials. Further, repeated measures analyses from 0 to 52 weeks and

0 to 2 years also found nonsignificant between-group differences. Arts et al. reported a repeated

measures AMD from 0 to 52 weeks of 0.7 (95% CI, -0.1 to 1.5) and from 0 to 2 years of 0.5

(95% IC, -0.3 to 1.3). See Appendix C, Table C-3 for detailed findings from Brouwer et al.,37

which are not directly comparable because of an adaptation to the standard scoring approach

used by this study (personal communication with author, February 8, 2018).

Other Pain Measures

Three RCTs reported on the frequency and proportion of participants with improvement in their

pain. These findings are largely consistent with previously reported pain outcomes for this



comparison; however, we did not use findings from these studies in the strength of evidence

ratings because of the heterogeneity in outcome definitions used. Specific study findings:

Mayer et al. reported the frequency and proportion of patients reporting low back pain

decreased from 20 (100%) at baseline to 4 (20%) at 2 years followup among participants

allocated to percutaneous endoscopic discectomy and from 20 (100%) to 7 (35%) among

participants allocated to microdiscectomy (calculated P=0.48).32 However, the proportion

reporting low back pain at 2 years was higher in the participants allocated to percutaneous

endoscopic discectomy (47.4% versus 20%, calculated P=0.18).

Ryang et al., the frequency and proportion of participants reporting radicular pain was 27

(90%) among those allocated to trocar discectomy and 29 (97%) among those allocated to

microdiscectomy, these frequencies and proportions decreased to 1 (3%) and 5 (17%) after

a mean followup of 1.3 years (P=0.11).28

Sasaoka et al. (N=24) reported the proportion of participants with residual low back pain or

lumbar discomfort at 52 weeks followup.24 Among participants allocated to

microendoscopic discectomy, 36.7% endorsed residual pain, among participants allocated

to microdiscectomy, 66.7% endorsed residual pain, the P value for comparison was

reported as not significant.

C. Microdiscectomy compared with discectomy

Three trials comparing microdiscectomy to discectomy reported pain outcomes using a VAS at 4

and 6 weeks,33 at 52 weeks,25 and at 26 weeks, 52 weeks, and 2 years.27 All three were rated as

some concerns for bias. Pain outcomes reported included VAS 10 cm measures for leg pain and

VAS 10 cm measures for back pain. Table 9 summarizes the findings and strength of evidence

related to pain outcomes for this comparison. A detailed description of findings follows this

table.

Table 9. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for pain in persons with symptomatic lumbar radiculopathy



№ of Studies

Risk of Bias


Pain (leg and back)-VAS 100 mm (short-term) (followup: range 4 weeks to 6 weeks) (MID 7 to 11 points)

1 RCT Seriousa Not seriousb Not serious Very seriousc

Pain decreased in both surgical groups, with no between-group differences. Henriksen et al.33 (N=80) only depicted outcomes on a figure, actual values and variance were NR.

⨁◯◯◯


Pain (leg and back)-VAS 10 cm (medium- and long-term) (followup: range 26 weeks to 2 years) (MID 7 to 11 points)

2 RCTs Seriousd Not serious Not serious Very seriouse

Improvements in both leg and back pain from baseline to followup in both surgical groups, however no between-group differences. Tullberg et al.25 (N=60) calculated 52w AMD 0.2 for leg pain, 0.1 for back pain (P Values NR). Teli et al. (N=142) calculated AMD 0 at 26w, 52w, and 2y for leg pain (P=0.73 for between-group differences). Similar pattern for back pain (P=0.75 for between-group differences).

⨁◯◯◯


a This RCT was rated as some concerns for bias. Sources of bias: inadequate randomization or allocation concealment and lack of

participant and outcome assessor blinding.



b Not applicable since body of evidence has only 1 study.

c Optimal information size criterion net met: the sample size in this study would be unable to detect differences less than

approximately 14 points, actual measured values and measures of variance not provided.

d Both studies were rated as some concerns for risk of bias. Sources of bias: inadequate randomization or allocation concealment

and lack of participant and outcome assessor blinding.

e Optimal information size criterion not met: the sample sizes of both studies would be unable to detect differences less than

approximately 16 points in one study25 and less than 10 points in the other.27; actual AMDs not reported and unable to calculate

confidence intervals based on data provided.

Abbreviations: AMD = absolute mean difference; MID = minimally important between-group difference; N = number; NR =

not reported; RCT = randomized controlled trial; VAS = visual analog scale; w = week(s); y = year(s)

In Henriksen et al., actual VAS values were not reported but no differences were reported

between groups for both VAS leg pain and VAS back pain at 4 weeks and at 6 weeks followup.33

Tullberg et al. reported a mean baseline VAS 10 cm leg pain score of 7.0 (SD NR) among

participants allocated to microdiscectomy and 7.0 (SD NR) among participants allocated to

discectomy.25 The mean scores at 52 weeks were 2.1 (SD NR) and 2.3 (SD NR), respectively

(AMDs and P value NR). The reduction in VAS 10 cm back pain scores was also similar in both

surgical groups (baseline 3.6 and 3.7, respectively, 52 weeks 1.6 and 1.8, respectively; AMDs

and P value NR). Teli et al. reported VAS 10 cm leg pain score of 8 (SD 1) at baseline

decreasing to 2 (1 SD) at 26 weeks, 1(1) at 52 weeks, and 2 (1) at 2 years in both surgical groups

(P=0.73 for between-group differences).27 A similar finding was observed for VAS 10 cm back

pain scores (P=0.75 for between-group differences).

Functioning/Disability


Five RCTs reported various measures of physical, mental, emotional, and social functioning or

disability.21,22,30,31,39 Three were rated as high risk of bias,21,30,31 one was rated as some concerns

for bias,22 and one was rated as high risk for outcomes at 12 weeks or later or some concerns for

outcomes at 6 weeks or later.39 Functional outcomes reported include the Oswestry disability

index, the physical functioning subscale of the SF-36, the Roland-Morris Disability

Questionnaire, the Prolo Scale, and various other subscales of the SF-36. We were unable to

conduct any quantitative synthesis for functional outcomes within this comparison because of

outcome heterogeneity because some studies did not report measures of variance needed to

conduct a meta-analysis. Table 10 summarizes the findings and strength of evidence related to

functioning/disability outcomes for this comparison. A detailed description of findings follows

this table.



Table 10. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for functioning/disability in persons with symptomatic lumbar radiculopathy (EQ1)


Summary of Findings CERTAINTY/Direction of Effect

№ of Studies

Risk of Bias


Functioning/Disability-Oswestry Disability Index (short- and medium-term) (followup: range 6 weeks to 26 weeks) (MID 5 to 17 points)


Not serious Not serious Seriousb Improvements over time in both treatment groups. Weinstein et al.[SPORT]21 (N=501) reported larger improvement in surgical group compared to conservative management group at 12w (AMD -4.7 [95% CI -9.3 to -0.2]). Osterman et al.31 (N=56) calculated 6w AMD -6, 12w AMD -6, 26w AMD -4 (P values NR).

⨁◯◯◯


Functioning/Disability-Oswestry disability index (long-term) (followup: range 52 weeks to 8 years) (MID 5 to 17 points)


Not serious Not serious Seriousc Improvements over time in both treatment groups; no significant between-group differences. Gerszten et al.39 (N=90) repeated measures AMD NR, P=0.08, from 0 to 52 weeks. Osterman et al.31 (N=56) repeated measures AMD -3 (95% CI, -10 to 4) from 0 to 2 years. Weinstein et al. [SPORT]21 (N=501) repeated measures AMDs from 0 to 2 years (AMD NR, P=0.21), 0 to 4 years (AMD NR, P=0.074), and 0 to 8 years (AMD NR, P=0.096).

⨁◯◯◯


Functioning/Disability-Roland-Morris Disability Questionnaire (short- and medium term) (followup: range 6 weeks to 26 weeks) (MID 2 to 5 points)


Not serious Not serious Seriouse Improvements over time in both treatment groups. Significant between-group difference in short term in 1 of the 2 trials. Peul et al.30 (N=283) 8w AMD -3.1 (95% CI, -4.3 to -1.7), 26w AMD -0.8 (95% CI, -2.1 to 0.5). McMorland et al.22 (N=40) repeated measures AMD at 12w NR, P=0.199.

⨁◯◯◯


Functioning/Disability-Roland-Morris Disability Questionnaire (long-term) (followup: range 52 weeks to 5 years) (MID 2 to 5 points)

1 RCT Very seriousf

Not seriousg Not serious Not serious Improvements over time in both treatment groups. Between group differences were not significant. Peul et al.30 (N=283) 52w AMD -0.4 (95% CI, -1.7 to 0.9), 2y AMD -0.5 (95% CI, -1.8 to 0.8), and 5y AMD 0.1 (95% CI, -1.3 to 1.4). Cumulative score 0 to 52w and 0 to 2y also with no significant difference between groups.

⨁⨁◯◯

LOW No difference

Functioning/Disability-SF-36 physical functioning subscale (short- and medium-term) (followup: range 6 weeks to 26 weeks) (MID 3 points)


Not serious Not serious Serioush Improvements over time in both treatment groups. Between-group differences favoring surgery in short-term in 1 of the 3 trials, otherwise no significant between-group differences. McMorland et al.22 (N=40) calculated AMDs 6w 1.2, 12w 11.5, repeated measure AMD at 12w NR, P=0.720). Weinstein et al.[SPORT]21 (N=501) AMD 12w 2.8 (95% C,I -2.5 to 8.1); Peul et al.30 (N=283) AMD 8w 9.3 (95% CI, 4.4 to 14.2), AMD 26w 1.5 [95% CI, -3.4 to 6.4]).

⨁◯◯◯


(continued)



Table 10. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for functioning/disability in persons with symptomatic lumbar radiculopathy (EQ1) (continued)



№ of Studies

Risk of Bias


Functioning/Disability-SF-36 physical functioning subscale (long-term) (followup: range 52 weeks to 8 years) (MID 3 points)


Not serious Not serious Seriousi Improvements over time in both treatment groups, but no significant difference between groups. Peul et al.30 (N=283) 52w AMD 2.2 (95% CI, -2.8 to 7.2). Weinstein et al.[SPORT]21 (N=501) no significant between-group differences at any single follow-up time; repeated measures AMD 0-2y (AMD NR, P=0.71), 0 to 4y (AMD NR, P=0.42), and 0 to 8y (AMD NR, P=0.47).

⨁◯◯◯


a Risk of Bias was high in all trials. Sources of bias: lack of participant and outcome assessor blinding and extensive and

differential crossovers.

b Optimal information size criterion not met in smaller trial:31 sample size able to detect differences larger than approximately

12.5 points.

c Optimal information size criterion not met in two smaller trials: sample size able to detect differences larger than approximately

12.5 points in 1 trial31 and approximately 9 in the other. AMDs and confidence intervals NR in the largest trial.21

d Risk of Bias was high in 1 trial,30 and was some concerns in the other trial.22 Sources of bias: lack of participant and outcome

assessor blinding in both trials, extensive crossovers in the trial rated as high risk of bias.

e Optimal information size criterion not met: the smaller trial did not report AMD or confidence intervals and sample size only

able to detect differences larger than approximately 4.5 points.

f Risk of Bias was rated as high in this trial. Sources of bias: lack of participant and outcome assessor blinding and extensive

crossovers.

g Not applicable as only 1 study is in this body of evidence.

h Optimal information size criterion not met: Sample size of 1,398 required for between-group difference of 3 points; smallest

trial22 unable to detect differences smaller than approximately 20 points, sample sizes of other trials unable to detect differences

smaller than 7 to 8 points.

i Optimal information size criterion not met: Sample size of 1,398 required for between-group difference of 3 points; sample size

of trials unable to detect differences smaller than 7 to 8 points. AMDs and confidence intervals NR in larger trial.21


difference; N = number; NR = not reported; RCT = randomized controlled trial; SF-36 = short-form 36 survey; SPORT = Spine

Patient Outcomes Research Trial; w = week(s); y = year(s)


Three RCTs reported outcomes using the Oswestry disability index.21,31,39 Higher scores on this

index represent worse functional status and a negative between-group difference favors surgery.

Across studies, function improved in both participants allocated to surgery and in participants

allocated to conservative management; between-group differences favor surgery in the short-

term, but differences do not persist in the long-term. Specific study findings:

Weinstein et al. [SPORT] reported this index at 12 weeks, 52 weeks, 2 years, 4 years, and 8

years.21,44,45 Only the AMD at 12 weeks demonstrated a significant difference between

treatment groups, favoring discectomy/microdiscectomy compared with conservative

management (AMD -4.7 [95% CI, -9.3 to -0.2]). All other follow-up time points

demonstrated a larger numeric improvement for participants allocated to surgery, but these



differences were not significant. Further, repeated measures from 0 to 2 years, 0 to 4 years,

and 0 to 8 years also did not demonstrate a significant AMD among participants allocated

to the discectomy/microdiscectomy and conservative management (AMDs NR, P=0.21,

P=0.074, and P=0.096, respectively).

Osterman et al. reported improvements in the index among participants allocated to

microdiscectomy and to participants allocated to physiotherapy at all follow-up time points

(6 weeks, 12 weeks, 26 weeks, 52 weeks, and 2 years).31 Although improvements were

numerically larger among participants allocated to microdiscectomy compared to

participants allocated to physiotherapy at all follow-up time points, the repeated measures

AMD from 0 to 2 years followup did not demonstrate a difference between treatment

groups (AMD -3 [95% CI, -10 to 4]).

Gerszten et al. reported a significant improvement in the index at 52 weeks in participants

allocated to plasma disc decompression and in participants allocated to epidural steroid

injection; however, no significant between-group differences were observed in the repeated

measures analysis (actual values NR, P=0.08).39


Two RCTs reported outcomes using the Roland-Morris Disability Questionnaire.22,30 Higher

scores represent worse functional status and a negative mean difference favors surgery. Across

studies, function improved in participants allocated to both surgery and conservative

management, and similar to the other function measures reported, function improved in some

studies in the short-term, but these differences did not persist in the long-term.

McMorland et al. reported significant improvements between baseline and 12 weeks in

participants allocated to microdiscectomy and in participants allocated to spinal

manipulation (AMDs NR, P=0.033).22 However, the repeated measures analysis found no

significant differences between treatment groups (AMDs NR, P=0.199).

Peul et al. also reported outcomes using this measure.30 At 8 weeks, a larger improvement

was observed in participants allocated to microdiscectomy compared with participants

allocated to conservative management (AMD -3.1 [95% CI, -4.3 to -1.7]). This difference

between groups did not persist at 26 weeks, 52 weeks, 2 years, or 5 years and cumulative

scores at 52 weeks and 2 years also demonstrated no significant difference between groups.

SF-36 Physical Functioning

Three RCTs reported outcomes using the physical functioning subscale of the SF-36.21,22,30

Higher scores on this index represent better functional status and a positive between-group

difference favors surgery. Across studies, function improved in participants allocated to both

surgery and conservative management. Some, but not all, studies reported significant between-

group differences favoring surgery in the short term, but these differences did not persist in the

long term. Specific study findings:

McMorland et al. reported improvements at 6 weeks and 12 weeks among participants

allocated to microdiscectomy and among participants allocated to spinal manipulation



(AMDs NR, P=0.034).22 However, repeated measures analysis found no significant

difference between treatment groups from baseline to 12 weeks (AMD NR, P=0.720).

Peul et al. also reported improvements in this scale over time among participants allocated

to microdiscectomy and in participants allocated to conservative management; the

between-group difference favored the surgical group at 8 weeks (AMD 9.3 [95% CI, 4.4 to

14.2]) but was not significantly different at 26 weeks (AMD 1.5 [95% CI, -3.4 to 6.4]) or

52 weeks (AMD 2.2 [95% CI, -2.8 to 7.2]).

Weinstein et al.[SPORT] reported slightly larger improvements in function at most follow-

up time points (range of AMD 0 to 2.8 at 12 weeks, 52 weeks, 2 years, 4 years, 8 years)

among participants allocated to discectomy/microdiscectomy compared to participants

allocated to conservative management, but these differences were not statistically

significant at any follow-up time point or in any of the three repeated measures analyses (0

to 2 years [AMD NR, P=0.71], 0 to 4 years [AMD NR, P=0.42], and 0 to 8 years [AMD

NR, P=0.47]).

Other Measures of Function

Two RCTs also reported on other various subscales of the SF-36, including social functioning,

mental functioning, role emotional, and role physical.22,30 Most scores improved over time

among participants allocated to surgical interventions and among participants allocated to

nonsurgical interventions but no significant differences in improvement between treatment

groups were observed. One RCT reported outcomes using the functional and economic subscales

of the Prolo scale.30 Peul observed a significant improvement in the functional score at 26 weeks

(AMD 0.5 [95% CI, 0.2 to 0.7]), but not at 8 weeks (AMD 0.8 [95% CI -0.6 to 1.1]) or 52 weeks

(AMD -0.04, [95% CI -0.3 to 0.2]). A similar pattern was observed for the economic subscale of

this measure.

B. Minimally invasive surgery compared to microdiscectomy or discectomy

Eight RCTs reported at least one outcome related to functioning or disability. One was rated as

low risk of bias,38 four were rated as some concerns for bias,26,27,34,37 and three were rated as high

risk of bias.28,29,40 Functional outcomes reported include the Oswestry Disability Index, the

Physical Functioning subscale of the SF-36, the Roland-Morris Disability Questionnaire, the

Prolo Scale, and other measures of function. Table 11 summarizes the findings and strength of

evidence related to functioning/disability outcomes for this comparison. A detailed description of

findings follows this table.



Table 11. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for functioning/disability in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Oswestry Disability Index (medium- and long-term) (followup: range 12 weeks to 2.8 years) (MID 5 to 17 points)


Not serious Not serious Seriousb Scores decrease (improve) over time in both surgical groups. No significant between-group differences at any follow-up time. Ruetten et al.29(N=200) calculated AMDs range -5 to -6 at 12w, 26w, 52w, 2y; Franke et al.34(N=100) RM AMD at 52w NR (P=0.08); Ryang et al.28(N=60) calculated AMDs range -1.98 to 3.6, P=0.83 for between-group difference at 1.3y, P reported as NS for between-group difference at 2.8y; Teli et al.27(N=142) calculated AMD range 1 to 2 at 26w, 52w, 2y.

⨁◯◯◯


Roland-Morris Disability Questionnaire (short-term) (followup: range 4 weeks to 8 weeks) (MID 2 to 5 points)

2 RCTs Not serious

Seriousc Not serious Seriousb Scores decreased (improved) over time in both surgical groups. Between group differences were significant at 4 weeks in Brouwer et al.37 (N=115)(AMD -2.5 (95% CI, -4.7 to -0.2) but not at 8w (AMD 0.1 [95% CI, -2.1 to 2.3). Between group differences reported by Arts et al.38 (N=328) were not significant (4w AMD 0.2 (95% CI, -1.1 to 1.4), 8w AMD 0.8 (95% CI, -0.4 to 2.1).

⨁⨁◯◯

LOW No difference

Roland-Morris Disability Questionnaire (medium- and long-term) (followup: range 26 weeks to 2 years) (MID 2 to 5 points)


Not serious Not serious Seriousb Scores decreased (function improved) persisted over time in both surgical groups. With one exception, between group differences were nonsignificant. Haines et al.40 (N=34) 26w calculated AMD 0.02 (P=0.74); Arts et al.38 (N=328) 26w AMD 1.0 (95% CI, -0.2 to 2.3), 52w AMD 1.3 (95% CI, 0.03 to 2.6), 2y AMD 0.8 (95% CI, -0.5 to 2.1), RM AMD 0 to 52w 0.8 (95 %CI -0.2 to 1.7), RM AMD 0 to 2y 0.6 (-0.3 to 1.6); Brouwer et al.37 (N=115) 26w AMD 2.2 (95% CI, -0.1 to 4.4), 52w AMD 1.1 (-1.1 to 3.4), 2y AMD -0.1 (95% CI, -2.4 to 2.2). RM AMD 0 to 52w 0.1 (95 %CI, -1.2 to 1.6); RM AMD 0 to 2y 0 (95% CI -1.3 to 1.3).

⨁◯◯◯


SF-36 Physical Functioning (short-term) (follow up: range 4 weeks to 8 weeks) (MID 3 points) 2 RCTs Not

serious Very seriousc Not serious Seriouse Scores increased (improved) over time in both

surgical groups. Between-group differences were not significant in Arts et al.38(N=328) 4w AMD -1.1 (95% CI, -5.6 to 3.3); 8w AMD -3.3 (95% CI, -7.8 to 1.1). However, Brouwer et al.37(N=115) between-group differences favored minimally-invasive surgery at 4w (AMD 18.4 [95% CI, 10.0 to 26.8]) but no differences at 8w (AMD 5.6 [95% CI, -2.7 to 13.9]).

⨁◯◯◯

VERY LOW Mixed findingsc

(continued)



Table 11. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for functioning/disability in persons with symptomatic lumbar radiculopathy (EQ1) (continued)



№ of Studies

Risk of Bias


SF-36 Physical Functioning (medium- term) (follow up: range 12 weeks to 26 weeks) (MID 3 points) 4 RCTs Very

seriousf Not serious Not serious Seriousg Improvements in scores persisted over time in both

surgical groups. No significant between-group differences in the medium-term; pooled between-group mean difference in scores at 12w to 26w -2.4 (95% CI, -6.1 to 1.2, 4 RCTs, 527 participants, I2=0%).26,37,38,40

⨁◯◯◯


SF-36 Physical Functioning (long-term) (follow up: range 52 weeks to 2 years) (MID 3 points) 4 RCTs Very

serioush Seriousk Not serious Seriousl Improvements in scores persisted over time in both

groups, but mixed findings for between-group differences. Ryang et al.28(N=60) calculated AMD at 1.3y -10.1 (P=0.64) and calculated AMD at 2.8y 6.0 (P=0.436). Thome et al.26(N=84) calculated AMD at 2y 7.0 (P=0.026), favoring minimally-invasive surgery. Arts et al.38(N=328) AMD 52w -4.8 (95% CI, -9.3 to -0.2) favoring microdiscectomy, but between-group differences at 2y and in RM 0 to 52w and 0 to 2y were NS. Brouwer et al.37(N=115) AMDs at 52w and 2y were NS, RM AMD 0 to 52w 4.3 (95% CI, -4.5 to 13.2), but RM AMD 0 to 2y favored minimally-invasive surgery (6.1 [95% CI, 0.5 to 11.7]).

⨁◯◯◯

VERY LOW Mixed findings

Prolo Scale (followup: range 4 weeks to 8 weeks) (MID 0.4 points)

2 RCTs Not serious

Seriousc Seriousm Not serious Scores increased (function improved) in both surgical groups. Arts et al.38 (N=328) between group differences were not significant in either the functional subscale (4w AMD 0 [95% CI, -0.3 to 0.2]; 8w AMD -0.1 [95% CI, -0.3 to 0.2]) or the economic subscale (4w AMD 0.2 [95% CI, -0.1 to 0.5]; 8w AMD 0.1[95% CI, -0.2 to 0.4]). Brouwer et al. significant between-group difference in economic subscale at 4w (AMD 1.1 [95% CI, 0.5 to 1.6]) but not at 8w (AMD 0.2 [95% CI, -0.3 to 0.8)] or in functional subscale (4w AMD 0.2 [95% CI, -0.2 to 0.6]; 8w AMD -0.2 [95% CI 0.6 to 0.3]).

⨁⨁◯◯

LOW No difference

Prolo Scale (medium- and long-term) (followup: range 12 weeks to 2 years) (MID 0.4 points)

3 RCTs Seriousn Not serious Seriousm Not serious Score increases (functional improvements) persisted in both surgical groups in Arts et al.38 (N=328) and Brouwer et al.37 (N=115) with no significant between-group differences at 26w or 52w or in repeated measures from 0 to 52w. No significant between-group differences observed at 12w to 26w in Thome et al.26 (N=84) in proportion of reporting sum of functional and economic subscale scores > 7 (0 [worse] to 10 [best] on a modified Prolo scale); 92% among participants allocated to sequestrectomy compared to 76% among participants allocated to discectomy (P=0.11).

⨁⨁◯◯

LOW No difference

a Two trials were rated as high risk of bias,29,105 2 trials were rated as some concerns.27,34 Sources of bias: lack of blinding

intervention and outcome assessors (all trials), and inadequate randomization and allocation concealment.29,105



b Optimal information size criterion not met: trials did not have a sample size to detect a between-group difference of at least 2

points (for Roland-Morris Disability Questionnaire) or 5 points (for Oswestry Disability Index), which is the low end of the

minimally important clinical difference for these measures.

c Inconsistent finding at 4w; much larger treatment effect in Brouwer et al.37 favoring minimally-invasive surgery; this finding is

not consistent with the findings at 8w or with the findings at 4w and 8w in the other RCT (Arts et al.38). For the SF-36 Physical

Functioning measure, the between-group difference was large enough that we considered the inconsistency a very serious

concern and led us to conclude overall mixed findings.

d One trial was low risk of bias,38 1 trial was some concerns for bias,37 and 1 trial was high risk of bias.40 Sources of bias: Sources

of bias: lack of blinding intervention and outcome assessors37,40 and inadequate randomization and allocation concealment and

high attrition.40

e Optimal information size criterion not met: sample size of 1,398 required to detect a difference of 3 points; sample size in Arts

et al.38 unable to detect differences smaller than 6 points and Brouwer et al37 unable to detect differences smaller than

approximately 11 points.

f Risk of bias was low in 1 trial38, some concerns in 2 trials26,37 and high in 1 trial.40 Sources of bias: lack of participant and

outcome assessor blinding in all but the low risk of bias trial, and inadequate randomization and high attrition.40

g Optimal information size criterion not met; a sample size of 1,398 required to detect a difference of 3 points, the pooled sample

size was only 527 participants; the largest trial sample size only able to detect differences of approximately 6 points or greater38,

and the smallest trial40 sample size only able to detect differences of approximately 15 points or greater.

h Risk of bias was low in 1 trial38, some concerns in 2 trials26,37 and high in 1 trial.28 Sources of bias: lack of participant and

outcome assessor blinding in all but the low risk of bias trial, and inadequate randomization and high attrition.28

k Some inconsistency in magnitude and direction of effects across the 4 studies.

l Optimal information size criterion not met: sample size of 1,398 required to detect a difference of 3 points; largest trial sample

size (Arts et al.38) unable to detect differences smaller than 6 points; smallest trial sample size (Ryang et al.28 unable to detect

differences smaller than approximately 12.5 points.

m This scale is based on an observers’ assessment of the patients’ functional status; thus, is less direct than other measures based

on patient-report.

n The risk of bias was low in 1 trial38, and some concerns in the other 2 trials.26,37 Sources of bias lack of participant and outcome

assessor blinding.

Abbreviations: AMD = absolute mean difference; CI = confidence interval; N = number; NS = not significant; NR = not

reported; RCT = randomized controlled trial; RM = repeated measure; SF-36 = short-form 36 survey; w = week(s); y = year(s)


Four RCTs reported outcomes using the Oswestry Disability Index.27-29,34 Higher scores on this

index represent worse functional status and a negative between-group difference favors

minimally invasive surgery. Across studies, function improved in both participants allocated to

minimally invasive surgery and in participants allocated to standard surgery; no significant

between-group differences were observed in the medium or long-term. Specific study findings:

Ruetten et al. reported that scores at 12 weeks decreased from baseline by 53 points among

participants allocated to endoscopic discectomy and by 47 points among participants

allocated to microdiscectomy; however, between group differences were reported as not

significant (calculated AMD -6).29 These decreases persisted at 26 weeks, 52 weeks, and 2

years (calculated AMDs range -5 to -6, all results reported as not significant).

Teli et al. reported decreases from baseline to 26 weeks of 28 points among participants

allocated to microendoscopic discectomy, 29 points among participants allocated to

microdiscectomy, and 27 points among participants allocated to discectomy; between-

group differences were not significant (calculated AMDs range 1 to 2).27Teli et al. reported

persistence of these decreases at 52 weeks (calculated AMD 2) and 2 years (calculated

AMD 1), with between-group differences remaining nonsignificant.



Franke et al. reported decreased scores among both groups, but no significant between-

group difference (P=0.08) at 52 weeks or 2 years (AMDs NR and unable to be

calculated).34

Ryang et al. reported decreases of 41 points among participants allocated to trocar

microdiscectomy and 44.7 points among participants allocated to microdiscectomy at 1.3

years (P=0.83, calculated AMD 3.6).28 Although scores increased slightly at 2.8 years

among participants allocated to microdiscectomy, the between-group differences remained

nonsignificant (calculated AMD -1.98).

Roland Morris Disability Questionnaire

Three RCTs reported outcomes with the Roland-Morris Disability Questionnaire.37,38,40 Higher

scores represent worse functional status and a negative mean difference favors minimally

invasive surgery. Across studies, function improved in participants allocated to both minimally

invasive surgery and standard surgery; between-group differences favoring minimally invasive

surgery at 4 weeks was observed by one study, this difference did not persist at 8 weeks and with

one exception, no other between-group differences were observed at other time points. Specific

study findings:

Arts et al. observed an AMD of 0.2 (95% CI, -1.1 to 1.4) at 4 weeks and 0.8 (95% CI, -0.4

to 2.1) at 8 weeks.38 Nonsignificant between-group differences persisted at 26 weeks

(AMD 1.0 [95% CI -0.2 to 2.3]). For long-term outcomes, a very small, significant

between-group difference was observed at 52 weeks (AMD 1.3 (95% CI 0.03 to 2.6) but

not at 2 years (AMD 0.8 (95 % CI -0.5 to 2.1).38 Repeated measures between-group

differences from 0 to 52 weeks (P=0.11) and 0 to 2 years (P=0.17) were not significant.

Brouwer et al. observed a significant between-group difference at 4 weeks (AMD -2.5

[95% CI, -4.7 to -0.2]) but not at 8 weeks (AMD 0.1 [95% CI, -2.1 to 2.3]).37

Nonsignificant between-group differences persisted at 26 weeks (AMD 2.2 [95% CI, -0.1

to 4.4]), 52 weeks (AMD 1.1 (95% CI, -1.1 to 3.4) ,and 2 years -0.1 (95% CI, -2.4 to 2.2).

In addition, nonsignificant repeated measures between-group differences were observed for

0 to 52 weeks and for 0 to 2 years.

Haines et al. reported no significant between-group difference (P=0.74) in change from

baseline scores at 26 weeks (calculated AMD 0.02).40

SF-36 Physical Functioning

Five RCTs reported outcomes with the SF-36 physical functioning subscale.26,28,38,40 Higher

scores on this index represent better functional status and a positive between-group difference

favors minimally invasive surgery. Function improved in participants allocated to both surgical

groups. Similar to the Roland-Morris Disability outcome previously reported for this

comparison, between-group differences favoring minimally invasive surgery at 4 weeks were

observed by one study.37 This difference did not persist at 8 weeks. In the medium-term (12

weeks to 26 weeks), the pooled between-group mean difference was -2.4 (95 % CI, -6.1 to 1.2),

4 RCTs, 527 participants, I2=0.0%, Appendix F, Figure-F4). Some between-group differences

were observed at 52 weeks and 2 years; but the findings were mixed with respect to which group

was favored.



Arts et al. and Brower et al. reported increases in scores from baseline to 4 weeks and 8

weeks among participants allocated to the minimally invasive surgical interventions and

among participants allocated to microdiscectomy.37,38 Arts et al. reported no between-group

differences at 4 weeks (AMD -1.1 (95% CI, -5.6 to 3.3) or 8 weeks (AMD -3.3 (95% CI -

7.8 to 1.1).38 In contrast, Brouwer et al. observed a significant, and clinically-relevant

between group difference favoring percutaneous laser disc decompression compared with

microdiscectomy at 4 weeks (AMD 18.4 [95% CI, 10.0 to 26.8]); this difference did not

persist at 8 weeks (AMD 5.6 [95% CI, -2.7 to 13.9]).37 No between group differences in the

medium-term were observed by Arts et al. (AMD at 26 weeks -3.9 [95% CI, -8.3 to 0.6]),38

or Brouwer et al. (AMD at 26 weeks -3.2 [95% CI, -11.6 to 5.1]).37 In the long-term,

Brouwer et al.37 observed an AMD of -3.2 [95% CI -11.6 to 5.2] at 52 weeks and 4.3 [95%

CI -4.5 to 13.2] at 2 years. Repeated measures AMD was 5.3 (95% CI, -0.7 to 11.2) for 0 to

52 weeks and was 6.1 [95% CI, 0.5 to 11.7]) at 0 to 2 years, favoring minimally-invasive

surgery. In contrast, Arts et al. reported a significant between-group difference favoring

microdiscectomy at 52 weeks (AMD -4.8 [95% CI, -9.3 to -0.2]), but this difference was

not present at 2 years (AMD 0.8 [95% CI, -0.5 to 2.1]).38 Repeated measures between

group differences from 0 to 52 weeks (-3.1 [95% CI, -6.8 to 0.7]) and 0 to 2 years (-2.8

[95% CI, -6.5 to 0.9])) were also not significant in Arts et al.

Thome et al. reported no significant between-group differences at 12 to 26 weeks

(calculated AMD 2.2, P=0.32 for between-group difference at followup).26 However,

between-group differences observed at 2 years were significant and favored minimally-

invasive surgery (calculated AMD 7, P=0.026 for between-group difference at followup).

Haines et al. reported no significant between-group differences at 26 weeks (calculated

AMD 2.9, P=0.96).40

Ryang et al. reported no significant between-group differences at 1.3 years (calculated

AMD -10.1, reported P for between-group comparison at followup=0.64) or at 2.8 years

(calculated AMD 6.0, reported P for between-group comparison at followup=0.436).28

Prolo Scale

Three RCTs reported functioning/disability using the Prolo Scale, which has a functional

subscale and an economic subscale.26,37,38 On this measure, higher scores represent better

function and a positive AMD favors minimally invasive surgery. Function improved in both

surgical groups; with one exception at 4 weeks in one study, between-group differences were not

significant at any time point.

Brouwer et al. reported no significant between-group differences at 4 weeks or 8 weeks in

the functional subscale at 4 or 8 weeks (AMD 0.2 [95% CI, -0.2 to 0.6]; AMD -0.2 [95%

CI -0.6 to 0.3], respectively) or in the economic scale at 8 weeks (AMD 0.2 [95% CI, -0.3

to 0.8), but did observe a significant difference in the economic subscale at 4 weeks (AMD

1.1 [95% CI, 0.5 to 1.6]).37 Arts et al. reported no significant between-group differences in

either subscale at 4 weeks or at 8 weeks.38 At 26 weeks and at 52 weeks, no significant

between-group differences were observed in either study and repeated measures between-

group differences from 0 to 52 weeks were also not significant. Thome et al. also reported

outcomes at 12 to 26 weeks and at 2 years using the Prolo scale, but used a modified scale



that rated each subscale using a 5-point Likert scale (instead of a 4 point) and summed the

functional and economic subscales to obtain a total score (range of total score could vary

from 2 to 10, higher scores represent more improvement). The proportion of participants

with a total score greater than or equal to 7 at 12 to 26 weeks was 92% among participants

allocated to sequestrectomy and 76% among participants allocated to microdiscectomy

(P=0.11). This study also compared the proportion of participants with scores in the

following categories: 1 to 4, 5 to 6, 7 to 8, and 9 to 10. No difference in proportion of

participants in these categories was observed at 12 to 26 weeks (P=0.852) or at 2 years

(P=0.20).

Other Measures of Function

Two RCTs reported outcomes related to function/disability in the medium- and long-term using

various subscales of the SF-36 including role-emotional, role-physical, and social

functioning.26,28 With few exceptions, no between-group differences were observed at any

followup.


One RCT rated as some concerns for bias compared microdiscectomy to standard discectomy

and reported outcomes with the Oswestry disability index.27 Table 12 summarizes the findings

and strength of evidence related to functioning/disability outcomes for this comparison. Teli et

al. observed scores improve at 26 weeks, 52 weeks, and 2 years among participants allocated to

microdiscectomy and in participants allocated to discectomy. Among participants allocated to

microdiscectomy, score decreases from baseline (40 [SD 4]) ranged from 25 to 29 points; among

participants allocated to discectomy score decreases from baseline (39 [SD 4]) ranged from 24 to

27 points. No significant between-group differences were observed (P=0.81).

Table 12. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for functioning/disability in persons with symptomatic lumbar radiculopathy



№ of Studies

Risk of Bias


Functioning/disability- Oswestry Disability Index (medium- and long-term) (followup: range 26 weeks to 2 years) (MID 5 to 17 points)

1 RCT Seriousa Not seriousb Not serious Seriousc Improvements in scores at 26w, 52w, and 2y in both surgical groups, but no significant between-group differences. Teli et al.27 (N=142) score decreases from baseline (40 [SD 4]) ranged from 25 to 29 points among participants allocated to microdiscectomy and score decreases from baseline (39 [SD 4]) ranged from 24 to 27 points among participants allocated to discectomy. No significant between-group differences were observed (calculated AMDs 26w 2, 52w 2, 2y 1, P=0.81 across comparisons at all time points).

⨁⨁◯◯

LOW No difference

a This RCT was rated as some concerns for bias. Sources of bias: lack of participant and outcome assessor blinding.


c This study’s sample size would be unable to detect differences less than approximately 8 points.

Abbreviations: AMD = absolute mean difference; MID = minimally important between-group difference; RCT = randomized

controlled trial; SD = standard deviation; w = week(s); y = year(s)



Quality of life


Two RCTs reported health-related QOL outcomes. One was rated as some concerns for bias22

and one was rated as high risk of bias.31 These studies reported outcomes using the total SF-36

score22 (sum of all normed subscales, possible range 0 to 800) and the 15D QOL measure (range

0 to 1.0).31 For both measures, a higher score represents better QOL and a positive between-

group difference favors surgery. Table 13 summarizes the findings and strength of evidence

related to QOL outcomes for this comparison. A detailed description of findings follows this

table.

Table 13. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for quality of life in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Quality of Life-Total SF-36 score (short-term) (followup: range 6 weeks to 12 weeks)

1 RCT Not serious

Not seriousb Not serious Very seriousc

Improvements over time in both treatment groups. McMorland et al.22 (N=40) observed no between-group differences in repeated measures AMD for cumulative total SF-36 score at 12 weeks (AMD NR, P=0.382).

⨁◯◯◯


Quality of life- 15D HRQOL (short, medium- and long-term) (followup: range 6 weeks to 2 years)

1 RCT Very seriousd

Not seriousb Not serious Seriouse Improvements over time in both treatment groups. Osterman et al.31(N=56) calculated between-group AMDs ranged from 0.01 to 0.05 at the various timepoints reported. Repeated measures AMD from 0 to 2 years was 0.03 (95% CI -0.01 to 0.07).

⨁◯◯◯


a Risk of bias was some concerns. Sources of bias: lack of participant and outcome assessor blinding.

b Not applicable as only 1 study is in this body of evidence.

c A sample size of 518 would be required to detect a minimum between-group difference of 10% of the value of baseline scores

(approximately 37 points for total SF-36 score).

d Risk of bias was high. Sources of bias: lack of participant and outcome assessor blinding and extensive and differential

crossovers.

e A sample size of 24 is required to detect a minimum between-group difference of 10% of the value of baseline scores

(approximately 0.08 for 15D HRQOL measure). A sample size of 75 is required to detect a significant between-group difference

of 0.03.

Abbreviations: MID = minimally important between-group difference; AMD = absolute mean difference; CI = confidence

interval; HRQOL = health-related quality of life; N = number; NR = not reported; RCT = randomized controlled trial; SF-36 =

short-form survey 36; w = week(s); y = year(s).

In both studies, QOL improved from baseline to followup in the surgery and nonsurgical

comparator groups; no significant differences between groups were observed by either study.

McMorland et al. reported a mean (SD) total SF-36 score of 379.5 (149.8) in participants

allocated to microdiscectomy and 381.3 (161.9) among participants allocated to spinal

manipulation at baseline. These scores improved to 429.1 (157.3) and 445.6 (142.8) at 6

weeks and 500.3 (179.7) and 484.6 (148.9) at 12 weeks, respectively. The repeated measures

AMD from 0 to 12 weeks found no between-group difference (AMD NR, P=0.382).22



Osterman et al. reported the 15D QOL measure at baseline and at various follow-up times

from 6 weeks to 2 years.31 Baseline QOL was 0.83 (SD 0.07) among participants allocated to

microdiscectomy and 0.84 (SD 0.06) among participants allocated to physiotherapy.

Calculated AMDs between groups were small ranging from 0.01 to 0.05 at 6 weeks, 12

weeks, 26 weeks, 52 weeks, and 2 years (P values NR). The repeated measures AMD in

scores over 0 to 2 years was -0.03 (95% CI, -0.07 to 0.01).


Three RCTs reported health-related QOL. Two RCTs were rated as some concerns for bias26,27

and one was rated as high risk of bias.28 Table 14 summarizes the findings and strength of

evidence related to QOL outcomes for this comparison. A detailed description of findings

follows this table.

Table 14. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for quality of life in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Quality of Life -SF-36 Physical Health and Mental Health Component Summary (followup: range 12 weeks to 3 years) (MID 2 points for PCS and 3 points for MCS)


Not serious Not serious Seriousb Scores increase (improvement in QOL) in both surgical groups from baseline over time. With one exception, no between-group differences observed. Ryang et al.28 (N=60) SF-36 MCS at 1.3y 51.9 (SD 7.8) vs. 44 (SD 13.2), P=0.03; no difference at 2.8y and no between-group difference in PCS. Teli et al.27 (N=142) no between-group differences in PCS (P=0.68) and MCS (P=0.78) across time at 26w, 52w, and 2y. Thome et al.26 (N=84) at 12w to 26w: PCS 43.6 (SD 9.7) vs. 41.5 (SD 10.7), P=0.41; MCS 53.6 (SD 9.8) vs. 50.6 (SD 12.0), P=0.26.

⨁◯◯◯


a 1 trial was rated as high risk of bias28 and two were rated as having some concerns for bias.26,27Source of bias: lack of

participant and outcome assessor blinding in all trials; inadequate randomization and allocation concealment in 1 trial.28

b Optimal information size criterion not met: the sample size of the largest RCT is unable to detect differences less than about 4

points, the others are unable to detect differences of less than about 6 points. Actual AMDs with confidence intervals NR.

Abbreviations: MID = minimally important difference; MCS = mental health component summary score; N = number; PCS =

physical health component summary score; QOL = quality of life; RCT = randomized controlled trial; SD = standard deviation;

w = week(s); y = year(s)

Three RCTs reported SF-36 physical health (PCS) and mental health (MCS) component

summary scores over 12 weeks to 2.8 years.26-28 In all studies, quality of life as measured by both

component scores improved over time in both intervention groups, and with one exception, no

statistically significant between-group differences were observed.

Ryang et al. observed a significant difference in the SF-MCS at the 1.3 year followup;

participants allocated to microdiscectomy had a higher score (mean 51.9 [SD 7.8])

compared with participants allocated to minimal access trocar microdiscectomy (mean 44.0

[SD 13.2], P=0.03), but it is not clear whether this comparison adjusted for small

differences in baseline scores.28 No significant difference in this score was observed at the

2.8-year followup (mean 48.8 [SD 10.5] and 48.4 [SD 9.4), P=0.892).



Teli et al. reported no significant between groups differences in the PCS (P=0.68) or MCS

(P=0.78) across time at 26 weeks, 52 weeks, and 2 years.27

Thome et al. reported no significant difference between groups in the PCS (P=0.41) and the

MCS (P=0.26) at 12 to 26 weeks followup.26


One RCT rated as some concerns for bias compared microdiscectomy to standard discectomy

and reported outcomes with the SF-36 physical health component summary (PCS) score and the

mental health component summary (MCS) score at 26 weeks, 52 weeks, and 2 years.27 Table 15

summarizes the findings and strength of evidence related to quality of life for this comparison.

For PCS, Teli et al. reported increases from baseline (21 [SD 4]) ranging from 19 to 23 points

among participants allocated to microdiscectomy at the various follow-up time points compared

with increases from baseline (22 [SD 4]) ranging from 18 to 22 points among participants

allocated to discectomy. No significant between-group differences were observed (P=0.68).

Similar findings were reported for the MCS (P=0.78 for between-group differences).

Table 15. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for quality of life in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Quality of life- SF-36 physical and mental health component summary (medium- and long-term) (followup: range 26 weeks to 2 years) (MID 2 points for PCS and 3 points for MCS)

1 RCT Seriousa not seriousb Not serious Seriousc Improvements at 26w, 52w, and 2y in both surgical groups, but no significant between-group differences. Teli et al.27 (N=142 analyzed) PCS increases from baseline (21 [SD 4]) ranging from 19 to 24 points among participants allocated to microdiscectomy compared with increases from baseline (22 [SD 4]) ranging from 16 to 22 points among participants allocated to discectomy. No significant between-group differences were observed (calculated AMD 26w 1, 52w 2, 2y 3, P=0.68 for between-group differences across time). Similar findings were reported for the MCS (calculated AMD 26w 2, 52w 0, 2y 2, P=0.78 for between-group differences across time).

⨁⨁◯◯

LOW No difference

a This RCT was rated as some concerns for bias. Sources of bias: lack of participant and outcome assessor blinding.


c Optimal information size criterion not met: the sample size in this study is only able to detect differences of about 4 points.

Abbreviations: AMD = absolute mean difference; MCS = mental health component summary score; MID = minimally

important difference; N = number; PCS = physical health component summary score; RCT = randomized controlled trial; SD =

standard deviation; w = week(s); y = year(s)



Two RCTs reported outcomes related to neurological symptoms, specifically sensory or motor

deficits.31,39 One was rated as high risk of bias31 and was rated as some concerns for bias for 6-

week outcomes and high risk of bias for outcomes at 12 weeks and later.39 Table 16 summarizes



the findings and strength of evidence related to neurological symptoms for this comparison. A

detailed description of findings follows this table.

Table 16. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for neurologic symptoms in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Neurologic symptoms (short- and medium-term)

2 Seriousa Not serious Not serious Very seriousb

Sensory and motor deficits improved over time in both treatment groups; no difference between groups. Gerszten et al.39 (N=90) no difference in proportion of participants with motor or sensory deficits between plasma disc decompression and epidural steroid injection at 6w; Osterman et al.31 (N=56) no difference in proportion of participants with muscle weakness at 6w (53.8% vs. 46.2%), 12w (42.3% vs. 46.2%), and 52w (28.6% vs. 30%).

⨁◯◯◯


a Risk of bias was high in 1 trial, and there were some concerns in the other trial.

b Both trials were underpowered to detect differences in proportions less than 35% (N=56) and 25% (N=90).

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial; w = week(s)

Gerszten et al. evaluated neurologic symptoms on each side and at each lumbosacral nerve root

level (i.e., 8 comparisons for muscle strength and 8 comparisons for tactile sensitivity); no

significant differences between groups (plasma disc decompression vs. epidural steroid injection)

were observed in all but one of these 16 comparisons at 6 weeks followup (actual values NR, P

value NR).39 Osterman et al. reported a similar proportion of participants with muscle weakness

among those allocated to microdiscectomy (N=28) compared with those allocated to

physiotherapy (N=28), respectively, at 6 weeks (53.8% vs. 46.2%), 12 weeks (42.3% vs. 46.2%),

and 52 weeks (28.6% vs. 30%). This study did not report between-group statistical significance

tests.31


Six RCTs comparing minimally invasive surgery to microdiscectomy26,28,29,32,34 and discectomy41

reported outcomes related to neurological symptoms. Three RCTs were rated as some concerns

for bias26,34,41 and three were rated as high risk of bias.28,29,32 Table 17 summarizes the findings

and strength of evidence related to neurological outcomes for this comparison. A detailed

description of findings follows this table.



Table 17. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for neurologic symptoms in persons with symptomatic lumbar radiculopathy (EQ1)


Impact CERTAINTY/ Direction of Effect

№ of Studies

Risk of Bias


Neurological symptoms (medium and long-term) (followup: range 12 weeks to 2 years)


Not serious Not serious Very seriousb

Over 12 w to 2y, neurological symptoms improved in both surgical groups from preoperatively to postoperatively with no significant difference between surgical groups in five of the RCTs;26,28,29,32,41 one RCT did not report findings by group.34

⨁◯◯◯


a Three RCTS were rated as some concerns for bias26,34,41 and three were rated as high risk of bias.28,29,32 Sources of bias: lack of

participant and outcome assessor blinding (all trials), inadequate randomization and allocation concealment.28,29,32

b Optimal information size criterion not met: measures of variance not provided in many studies, unclear what effect size is

meaningful and whether most studies had adequate sample sizes to detect a minimally important difference.

Abbreviations: RCT = randomized controlled trial; w = week(s); y = year(s)

Findings were not reported by group in one RCT,34 the remaining five studies observed no

between-group differences. Three RCTs reported no statistical difference in neurological

symptoms between intervention groups26,28,29 We calculated no differences in the other 2

RCTs.32,41

Ruetten et al. reported mean North American Spine Society Neurology Scores ranging

from 1.9 to 2.1 over 12 weeks to 2 years followup in participants allocated to endoscopic

discectomy and mean scores ranging from 1.7 to 2.3 among participants allocated to

microdiscectomy; differences between groups were reported as not statistically significant

(P value NR).29

Ryang et al. reported no difference in the proportion of participants with sensory deficits

(40% vs. 43%) in participants allocated to minimal access trocar discectomy compared

with participants allocated to microdiscectomy, respectively, over an average of 1.3 years

followup (P=0.31).28 Similar findings were observed for the proportion with motor deficits

(27% vs. 23%, P=0.86).

Thome et al. reported improvements in both sensory and motor deficits from baseline to 2

years in participants allocated to both sequestrectomy and microdiscectomy.26 However,

the study reported no between-group differences in a composite measure of neurologic

symptoms at 2 years (actual value NR, P=0.278) or in the proportion of participants with

sensory deficits (actual value NR, P=0.52) or motor deficits (actual value NR, P=0.74) at

12 to 26 weeks followup.26

Hermantin et al. reported the proportion of participants with postoperative sensory deficits

(53.3% vs. 60.0%, calculated P=0.79) and motor weakness (16.7% vs. 33%, calculated

P=0.23) in participants allocated to video-assisted arthroscopic microdiscectomy and those

allocated to discectomy respectively at an unspecified follow-up time.41

Mayer et al. reported the proportion of participants with sensory deficit (5% vs. 25%,

calculated P=0.18) and motor deficit (0% vs. 0%) at 2 years followup among those



allocated to percutaneous endoscopic discectomy and those allocated to microdiscectomy,

respectively.32

Franke et al. reported that overall 83% of motor deficits and 68% of sensory deficits were

resolved completely at 52 weeks; outcomes were not reported by intervention group.34


No studies reported outcomes related to neurological symptoms for this comparison.

Return to work


Five RCTs reported various outcomes related to “return to work.”5,21,31,35,39 Some measures

captured actual return to work, whereas others reflected somewhat indirect measures, such as

self-reported ability to work, receipt of disability benefits, or pain affecting occupational status.

We rated all RCTs as high risk of bias for this outcome. Table 18 summarizes the findings and

strength of evidence related to return to work outcomes for this comparison. A detailed


Table 18. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for outcomes related to return to work in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Return to work


Not serious Seriousb Seriousc With the exception of one study, no between-group differences in measures relating to return to work. Erginousakis et al.35 (N=62) pain affecting occupational status (12w 12.9% vs. 9.7%, calculated P=1.0; 52w 12.9% vs. 71%, calculated P< 0.001). Weber et al.5 (N=126) receiving disablement benefits (4y 5% vs. 12.1%, calculated P=0.21; 10y 11.7% vs. 12.1%, calculated P=1.0). Gerszten et al.39 (N=90) return to work (69% vs. 70%,). Osterman et al.31 (N=56) VAS 100 work ability RM 0 to 2y AMD 5 (95% CI, -7 to 18). Weinstein et al.[SPORT]21 (N=501) 2y AMD proportion working full time -2.2% (95% CI, -10.6% to 6.2%).

⨁◯◯◯


a Risk of bias was high in all trials. Sources of bias: lack of participant and outcome assessor blinding and inadequate in all trials

and inadequate randomization and allocation concealment35 or extensive and differential crossovers.5,21,31,39

b Measures used in some studies were indirect: Erginousakis et al.35 (i.e. pain affecting occupational status); Osterman et al.31

work ability based on self-reported VAS 100 mm; Weber et al.5 reported on proportion receiving permanent disability benefits.

c Optimal information size criterion not met: variation in measures used and lack of clarity with respect to minimally important

differences, all but one study21 have sample sizes that are too small to detect small to modest differences in dichotomous

outcomes.

Abbreviations: AMD = absolute mean difference; N = number; RCT = randomized controlled trial; RM = repeated measure;

SPORT = Spine Patient Outcomes Research Trial; VAS = visual analogue scale; w = week(s); y = year(s)



Return to work outcomes were measured at various follow-up times across this body of

evidence. With one exception35; no between-group differences in return to work outcomes were

observed.

Erginousakis et al. reported minimal difference in the proportion of participants reporting

pain that affected their occupational status at 12 weeks followup (12.9% percutaneous disc

decompression, 9.7% conservative management, calculated P=1.0).35 This proportion

remained the same among participants that received percutaneous disc decompression but

increased to 71% at 52 weeks and 2 years among participants that received conservative

therapy (calculated P<0.001 at both time points).

Weber et al. reported the proportion of participants with a permanent incapacitation and

receiving disablement benefits.5 At 4 years followup, this proportion was 5% in

participants allocated to discectomy compared with 12.1% in participants allocated to

conservative management (calculated P=0.21). At 10 years followup the proportions were

11.7% and 12.1%, respectively (calculated P=1.0).

Osterman et al. reported an increase in work ability as measured by a 0 to 100 VAS score

over five follow-up time points from 6 weeks to 2 years in participants allocated to

microdiscectomy and in participants allocated to physiotherapy.31 Although participants

allocated to microdiscectomy had higher VAS work ability scores at all time points,

particularly at 12 weeks and 26 weeks, the between-group AMD in repeated measures

analysis over 0 to 2 years was not significant (AMD 5 [95% CI, -7 to 18]).

Weinstein et al. [SPORT] reported in increase in the proportion of participants working full

or part time in participants allocated to discectomy and in participants allocated to

conservative management from baseline to all follow-up times.21 However, between-group

differences in the proportion working full-time were not significant at any time point (12

weeks AMD -5.6% [95% CI, -14.5% to 3.4%], 52 weeks AMD -0.6% [95% CI, -8.6% to

7.3%], 2 years AMD -2.2% [95% CI, -10.6% to 6.2%], 4 years AMD −3.8 [95% CI, −13.3

to 5.8]).

Gerszten et al. reported that the proportion of participants working full or part time at 26

weeks was similar (69% to 70%) among participants allocated to plasma disc

decompression compared with participants allocated to epidural steroid injection.39


Six RCTs reported various outcomes related to “return to work”, though in some studies this

outcome was not reported by group.26,29,32,34,36,41 Four RCTs were rated as having some concerns

for bias26,34,36,41 and two were rated as having high risk of bias.29,32 Table 19 summarizes the

findings and strength of evidence related to work outcomes for this comparison. A detailed




Table 19. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for return to work outcomes in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Return to Work



Mean duration of postoperative disability is lower by a range of 4w to 15 w among participants allocated to minimally invasive surgery compared to standard surgery. Hermantin et al.41 (N=60) 3.9w (SD NR) vs. 7w (SD NR), P value NR; Mayer et al.32 (N=40) 7.7w (range 1 to 26w) vs. 22.9w (range 4 to 52w), P value NR; Ruetten et al.29 (N=200) 3.57w (SD NR) vs. 7w (SD NR), P< 0.01. Thome et al.26(N=84) reported no between-group differences using a multi-level categorical outcome of “work impairment”. Two RCTs did not report findings by group.34,36

⨁◯◯◯

VERY LOW Favors minimally-invasive surgeryb

a Two trials were rated as high risk of bias29,32 and four were rated as some concerns.26,34,36,41 Sources of bias: lack of participant

and outcome assessor blinding in all trials and inadequate randomization and allocation concealment in the two trials rated as

high risk of bias.

b Three of the 4 RCTs that reported between-group differences evaluated return to work using a continuous measure of

postoperative work disability in week and reported a magnitude of effect that seems clinically important. The RCT that reported

no between-group differences in return to work used a 4-level categorical measure of work impairment. The difference in type of

measure may explain why no between-group differences were observed in this study and we did not consider this an inconsistent

finding. Thus, we assessed the overall direction of effect for this body of evidence as favoring minimally-invasive surgery.

c Optimal information size criterion not met: two studies did not report findings by group to allow for estimate of between-group

difference; three studies did not report measures of variance, unclear what a minimally important difference is for this outcome

and what sample size would be required to detect a small to modest difference.

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; w = week(s)

Of the 4 RCTs that reported between-group differences, three RCTs29,32,41 suggest that

participants allocated to minimally-invasive surgery return to work sooner than participants

allocated to standard surgery as measured by weeks of postoperative disability. The range of this

difference is 4 weeks to 15 weeks. The remaining RCT26 reported no significant between-group

differences; however; this study used a multi-level categorical measure of work impairment,

which may be measuring a related, but different construct compared to the other three RCTs.

Ruetten et al. reported the mean duration of postoperative work disability was 3.57 (SD

NR) weeks in participants allocated to endoscopic discectomy compared with 7 (SD NR)

weeks in participants allocated to microdiscectomy (P< 0.01).29

Hermantin et al.41 reported a mean duration of postoperative disability in time lost from

work or until able to resume normal activity of 3.9 weeks among participants allocated to

video-assisted arthroscopic microdiscectomy compared with 7 weeks among participants

allocated to discectomy (measures of variance and P values NR).

Mayer et al.32 reported a mean duration of postoperative disability of 7.7 weeks (range 1 to

26) among participants allocated to percutaneous endoscopic discectomy compared with

22.9 weeks (range 4 to 52) among participants allocated to microdiscectomy (measures of

variance and P values NR).



Thome et al. reported specific categories of impairment of work at 12 to 26 weeks and at 2

years.26 Thirty-one percent of participants allocated to sequestrectomy reported that their

work impairment was “much better” at 12 to 26 weeks compared with 33% of participants

allocated to microdiscectomy. At 2 years, the proportions were 37% and 31%, respectively.

The proportion of participants endorsing various categories of work impairment were not

significantly different between groups (P=0.415 at 12 to 26 weeks, P=0.112 at 2 years).

Chatterjee et al. reported that 92.5% of participants allocated to microdiscectomy returned

to work or their previous level of activity by 12 weeks followup.36 The number of

participants returning to work in the group allocated to automated percutaneous lumbar

discectomy was not reported.

Franke et al. reported the mean duration of postoperative inability to work overall was 7

weeks; this duration was not reported by intervention groups (microscopically assisted

percutaneous nucleotomy vs. microdiscectomy).34


One RCT rated as some concerns for bias compared microdiscectomy to discectomy and

reported on outcomes related to “return to work”.25 Tullberg et al. reported a mean duration of

postoperative, full-time sick leave of 10.4 (SD NR) weeks in participants allocated to

microdiscectomy compared with 10.1 (SD NR) weeks in participants allocated to discectomy (P

value NR). The proportion out of work at an unspecified follow-up time point was 16.7% among

those allocated to microdiscectomy and 6.7% among those allocated to discectomy (calculated P

=0.42). Table 20 summarizes the findings and strength of evidence related to return to work for

this comparison.

Table 20 Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for return to work outcomes in persons with symptomatic lumbar radiculopathy (EQ1)



№ of Studies

Risk of Bias


Return to work


No between-group differences in return to work outcomes. Tullberg et al.25 (N=60) mean duration of postoperative, full-time sick leave 10.4w (SD NR) vs. 10.1w (SD NR) (P value NR).The proportion out of work at an unspecified follow-up time point 16.7% vs. 6.7% (calculated P=0.42).

⨁◯◯◯


a This RCT was rated as some concerns for bias. Source of bias: lack of participant and outcome assessor blinding and inadequate

information to evaluate randomization and allocation concealment process.


c Optimal information size criterion not met: no measures of variance provided for continuous measure and a sample size of 322

would have been required to detect a significant between-group difference for the difference in proportion observed by the

study.

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; w = week(s)



Other Efficacy Outcomes


Four RCTs reported other efficacy outcomes related to perceived recovery, overall time to

recovery, overall result, and patient satisfaction with symptoms.31, #46, #1839, #33 These outcomes

were consistent with previously reported efficacy outcomes that suggest more favorable

outcomes for participants who are allocated to surgery in the short and medium term. Results

from three of these studies also suggest some favorable outcomes in the long-term. We did not

use these outcomes in our strength of evidence ratings because of heterogeneity in outcome

definition. Specific study findings:

Osterman et al. found a higher proportion of participants reporting full recovery at 6 weeks

(19.2% vs. 0%, P< 0.05) among participants allocated to microdiscectomy (N=28)

compared with participants allocated to physiotherapy (N=28).31 The difference in

proportion was similar at 12 weeks (19.2% vs. 15.4%, calculated P=1.0) and 52 weeks

(33.3% vs. 25%, calculated P=0.73).

Peul et al. reported a significant difference in median time to recovery (4.0 weeks (95% CI,

3.7 to 4.4) vs. 12.1 weeks (95% CI 9.5 to 14.9; AMD NR, P< 0.001) among participants

allocated to microdiscectomy compared with participants allocated to conservative

management.30 The relative difference in time to “complete” or “nearly complete” recovery

at 52 weeks favored microdiscectomy (hazard ratio 1.97 [95% CI 1.72 to 2.22]. However,

the proportion of participants reporting complete or nearly complete recovery (on a 7-point

Likert scale of self-perceived recovery) by 2 years was not different between groups

(81.3% vs. 83.6%, AMD -2.4% [95% -12.0% to 7.2%]).

Weber et al. reported the proportion of participants achieving good, fair, poor, or bad

results overall.5 This assessment was based on an outcome assessor’s evaluation of the

patient’s neurological deficits, working capacity, pain, and mobility of the lumbar spine. At

52 weeks, 65.0% of participants allocated to discectomy had achieved “good” results

compared with 36.4% of participants allocated to conservative management. The difference

in proportion between groups across all four categories of results (good, fair, poor, bad)

was significant (P=0.0015) at 52 weeks, but these differences did not persist at 4 years or

10 years (P reported as not significant for both).

Weinstein et al.[SPORT] reported on proportion of patients endorsing various categories of

satisfaction with symptoms.21 At 12 weeks, a higher proportion (54.3%) of participants

allocated to discectomy/microdiscectomy reported being very or somewhat satisfied with

symptoms compared with 43.0% of participants allocated to conservative management

(AMD 11.3% [95% CI 1.6% to 20.9%]). This proportion remained numerically higher at

all subsequent follow-up time points (52 weeks, 2 years, 4 years, 8 years) among

participants allocated to discectomy/microdiscectomy but the difference between groups

was not statistically significant at any single follow-up time point. Repeated measures of

this proportion over 0 to 2 years and 0 to 4 years also suggest no significant difference

between groups; however, repeated measures of this proportion over 0 to 8 years was

statistically significant (P=0.013).




Ten RCTs reported other efficacy outcomes, related to perceived recovery, overall time to

recovery, overall result, and patient satisfaction with symptoms. With few exceptions, most

observed no significant differences between groups. We did not use these outcomes in our

strength of evidence ratings because of heterogeneity in outcome definition. Specific study

findings:

Arts et al. reported the frequency and proportion of patients with “complete” or “nearly

complete” recovery based on a 7-point Likert scale.38 Fewer participants allocated to

tubular discectomy reported complete or nearly complete recovery at all time points

measures (4 weeks, 8 weeks, 26 weeks, 52 weeks, and 2 years); however, the differences

between groups were only significant at 8 weeks (63% vs. 75%, odds ratio (OR) 0.56 [95%

CI, 0.35 to 0.92]) and 52 weeks (69% vs. 79%, OR 0.59 [95% CI, 0.35 to 0.99]). No

relative difference between groups in time to recovery by 52 weeks was observed (hazard

ratio (HR) 0.92 [95% CI, 0.73 to 1.17]).

Brouwer et al used the same 7-point Likert scale and reported no significant difference in

the frequency and proportion achieving complete or nearly complete recover among

participants allocated to percutaneous laser disc decompression versus those allocated to

microdiscectomy (69% vs. 75% at 52 weeks [OR 0.81, 95% CI, 0.4 to 1.9]; 70.8% vs.

60.8% at 2 years [OR 1.6, 95% CI, 0.7 to 3.6]).37 Unlike Arts et al. which did not find a

relative difference in time to recovery between groups, Brouwer et al. observed a

significantly slower recovery among participants allocated to percutaneous laser disc

decompression (HR 0.64 [95% CI, 0.42 to 0.97] at 52 weeks and similar findings at 2

years.

Three RCTs reported outcomes using the MacNab criteria, a 4-point Likert scale that rates

the overall outcome as excellent, good, fair, or poor. Chatterjee et al. reported a significant

difference in participants achieving an excellent or good outcome among participants

allocated to automated percutaneous lumbar discectomy (9 [29%])) compared with

participants allocated to microdiscectomy (32 [80%], p< 0.001).36 Haines et al. reported 11

(64.7%) of participants allocated to automated or endoscopic percutaneous discectomy

achieved an excellent or good outcome compared with 6 (60%) allocated to discectomy

(P=0.81).40 Lastly, Huang et al. reported 9 (90%) of participants allocated to

microendoscopic discectomy reported an excellent or good outcome compared with 11

(91.6%) of participants allocated to discectomy (calculated P=1.0).23

Other studies used various measures of satisfaction with outcome, satisfaction with

surgery, or global outcome rating. Mayer et al. reported the frequency and proportion with

self-reported success of surgery at 2 years; 9 (47%) of participants allocated to

percutaneous endoscopic discectomy compared with 8 (40%) of participants allocated to

microdiscectomy.32 Hermantin et al. reported no difference in frequency and proportion

with satisfactory outcome between participants allocated to video-assisted arthroscopic

microdiscectomy (29 [97%]) compared with 28 [93%]) allocated to discectomy (calculated

P=1.0).41 Ryang et al. used a VAS 10 cm to report overall improvement from baseline to

2.8 years followup; although scores reflected overall improvement among participants

allocated to trocar discectomy and participants allocated to microdiscectomy, the scores at



followup were similar (4.92 vs. 4.64, P value NR). 28 Sasaoka et al. reported a mean

percentage improvement in the Japanese Orthopaedic Association Score at 52 weeks of

84.7% among participants allocated to microendoscopic discectomy compared with 88.6%

among participants allocated to microdiscectomy (P value NR).24 Thome et al. reported no

differences in patient satisfaction with surgery scores at 12 to 26 weeks or 2 years.26


One RCT rated as some concerns for bias comparing microdiscectomy to discectomy reported

the frequency and proportion of participants with a specified opinion on recovery at 52 weeks

(total recovery, almost recovered, good, unchanged, or worse).25 Among participants allocated to

microdiscectomy, 11 (37.9%) reported total recovery, and 8 (27.6%) reported almost recovered.

Among participants allocated to discectomy these outcomes were 6 (20.7%) and 14 (28.3%),

respectively (calculated P=0.25 and 0.18, respectively).

3.2.2 Efficacy Question 2

In adults with symptomatic lumbar radiculopathy, does effectiveness or comparative

effectiveness of surgical interventions vary for patients who are not employed because of

disability or patients who are undergoing recurrent surgery for relapse?

We did not identify any studies that reported outcomes specifically for patients not employed

because of disability. We identified two studies focused on the efficacy47 or comparative

effectiveness48 of revision surgery for relapse. Both were rated as high risk of bias. Study and

population characteristics of these trials are summarized in Table 21. The strength of evidence

ratings and summary of findings from these studies are provided in Table 22 and Table 23. A

detailed description of study characteristics and findings follows these tables. Appendix C,

Tables C-1 and C-2 provides evidence tables with individual study and population

characteristics. Appendix C, Tables C-3 and C-4 provide evidence tables with detailed individual

study outcomes related to efficacy.



Table 21. Study and population characteristics of the two randomized controlled trials comparing revision surgical interventions to spinal cord stimulation or an alternative revision surgery for the management of lumbar radiculopathy relapses (EQ2)





Primary Outcomea (effect size detectable with 80%

power, = 5%);

Other outcomes

North (2005)47 United States; High

Age: 52.0 (13.5) Women: 26 (52%) Mean (SD) number of prior operations: SG1: 2.5 (1.1) NS1: 2.5 (1.1)

Repeat lumbosacral decompression N randomized: 26; N analyzed: 26 (100%); N crossovers: 14 (54%)

Spinal cord stimulation N randomized: 24; N analyzed: 19 (79.2%); N crossovers: 5 (20.8%)

Successful treatment (NR);

Impairment from pain in performing everyday activities

Return to work

N (%) with successful treatment

Ruetten (2009)48 Germany; High

Age: 39 (range 23 to 59) Female: 44 (44%) Duration of symptoms, mean (range) in weeks: 9.85 (0.14 to 56)

Revision endoscopic discectomy N randomized: 50; N analyzed: 45 (90%); N crossovers: 0 (0%)

Revision microdiscectomy N randomized: 50; N analyzed: 42 (84%); N crossovers: 0 (0%)

NR (NR);

VAS 100 leg pain

VAS 100 back pain

NASS pain score

NASS neurology score


Postoperative work disability

N (%) with no leg pain

N (%) satisfied with surgery and would undergo the operation again


Abbreviations: N = Number; NASS = North American Spine Society; NR = not reported; NS = nonsurgical group; SD =

standard deviation; SG = surgical group; VAS = visual analogue scale



Table 22. Summary of findings for pain, functioning, neurological symptoms and quality of life for RCTs for repeat lumbosacral decompression surgery compared with spinal cord stimulation for treatment of lumbar radiculopathy relapses (EQ2)


Summary of Findings Certainty/Direction of Effect

№ of Studies

Risk of Bias


Pain (short-, medium-, and long-term)

0 RCT

-

Functioning/Disability-impairment in everyday activities due to pain (long-term) (followup: range 1.8 years to 5.7 years)

1 RCT Very seriousa

Not seriousb Seriousc Very seriousd

North et al.47(N=50) reported higher levels of impairment from pain in performing everyday activities among participants allocated to repeat surgery compared to participants allocated to spinal cord stimulation, values were only depicted on a figure and differences between groups were reported as NS.

⨁◯◯◯


Quality of life (short-, medium-, and long-term)

0 RCT

-

Neurologic symptoms (short-, medium-, long-term)

0 RCT

-

Return to work (long-term) (followup: range 1.8 years to 5.7 years)

1 RCT Very seriousa

Not seriousb Seriousc Very seriousd

North et al.47(N=50) reported no significant differences in return to work but actual values were NR.

⨁◯◯◯


a The trial was rated as high risk of bias. Sources of bias: lack of participant and outcome assessor blinding, extensive deviations

from intended interventions and differential attrition.

b Not applicable as this body of evidence has only one study.

c This measure was not well defined and seems to assess pain as well as functional impairment.

d Optimal information size criterion net met: actual values and measures of variance were not reported and study sample size is

unlikely to be able to detect anything but a very large difference.

Abbreviations: N= number; NR = not reported; NS = not significant; RCT = randomized controlled trial;



Table 23. Summary of findings for pain, functioning, quality of life, neurologic symptoms and return to work comparing revision endoscopic discectomy to revision microdiscectomy for treatment of lumbar radiculopathy relapses (EQ2)


Summary of Findings CERTAINTY/Direction of effect

№ of Studies

Risk of Bias


Pain (leg)-VAS 100 mm (short- medium- and long-term) (followup: range 12 weeks to 2 years) (MID 7 to 11 points)

1 RCT Very seriousa


Ruetten et al.48 (N=100) reported improvements over time from baseline to 12w, 26w, 52w, and 2y among participants in both surgical groups. Between-group differences were reported as NS.

⨁◯◯◯


Pain (back)-VAS 100 mm (short-, medium-, and long-term) (followup: range 12 weeks to 2 years) (MID 7 to 11 points)

1 RCT Very seriousa


Ruetten et al.48 (N=100) reported improvements over time from baseline to 12w, 26w, 52w, and 2y among participants in both surgical groups. Between-group differences were reported as NS.

⨁◯◯◯


Pain-North American Spine Society Pain Score (short-, medium- and long-term) (followup: range 12 weeks to 2 years)

1 RCT Very seriousa

Not seriousb Not serious Seriousd Ruetten et al.48 (N=100) reported improvements over time from baseline to 12w, 26w, 52w, and 2y among participants in both surgical groups. Between-group differences were reported as NS.

⨁◯◯◯


Functioning/Disability-impairment in activities due to pain (long-term) (followup: range 12 weeks to 2 years)

1 RCT Very seriousa

Not seriousb Not serious Seriousd Ruetten et al.48 (N=100) reported improvements as measured by the Oswestry Disability Index from baseline to 12w, 26w, 52w, and 2y among participants allocated to both surgical groups. Between-group differences were reported as NS.

⨁◯◯◯


Quality of life (short-, medium-, or long-term)

0 RCT

-

Neurological symptoms (medium- to long-term) (followup: range 12 weeks to 2 years)

1 RCT Very seriousa

Not seriousb Not serious Seriousd Ruetten et al.48 (N=100) reported improvements in both surgical groups over time in the mean North American Spine Society Neurology scores from baseline to 12w, 26w, 52w, and 2y. Between-group differences were reported as NS.

⨁◯◯◯


Return to work-mean duration of postoperative disability

1 RCT Very seriousa

Not seriousb Not serious Seriousd Ruetten et al.48 (N=100) reported a significant different in mean duration of postoperative disability (P < 0.01): 4w (SD NR) among participants allocated to revision endoscopic discectomy and 7.4w (SD NR) among participants allocated to revision microdiscectomy.

⨁◯◯◯

VERY LOW Favors revision endoscopic discectomy

a This trial was rated as high risk of bias. Sources of bias: lack of participant and outcome assessor blinding, inadequate

randomization and allocation concealment process.

b Not applicable as this body of evidence has only one study.

c Optimal information size criterion not met: measures of variance not reported and sample size only able to detect differences of

about 15 points.

d Optimal information size criterion not me: measures of variance were not reported, lack of clarity about minimally important

differences for these measures.

Abbreviations: MID = minimally important between-group difference; N= number; NR = not reported; NS = not significant;

RCT = randomized controlled trial; SD = standard deviation; VAS = visual analogue scale; w = week(s); y = year(s)



3.2.2.1 Study Characteristics

Both included studies were parallel-group RCTs. North et al. was conducted in the United States

and compared repeat lumbosacral decompression (N randomized=26) with spinal cord

stimulation (N randomized=24) in adults with persistent radicular pain despite one or more prior

lumbosacral spine surgeries.47 The mean number of prior surgeries was 2.5 in both surgical

groups, and overall 30% of participants reported receiving workers compensation benefits.

Ruetten et al. was conducted in Germany and compared revision endoscopic discectomy (N

randomized=50) to revision microdiscectomy (N randomized=50) in adults who had a previous

conventional discectomy with acute occurrence of radicular leg symptoms after a pain-free

interval in combination with a recurrent disc herniation on MRI.48 The mean duration of

symptoms overall in this study was 9.85 weeks (range 0.14 to 56). The proportion that were

disabled or receiving disability benefits was not reported. North et al. was rated high risk for bias

because of substantial crossover between groups, differential attrition and Ruetten et al. was

rated high risk for bias because randomization and allocation concealment were inadequate. In

addition, both studies did not blind participants or outcome assessors.

3.2.2.2 Findings

Pain

North et al. did not report any outcomes related to pain. Ruetten et al. reported improvement in

VAS 100 mm leg pain score from baseline to 12 weeks, 26 weeks, 52 weeks, and 2 years among

participants allocated to revision endoscopic discectomy and among participants allocated to

revision microdiscectomy.48 Between-group differences were reported as not significant at any

follow-up time point (AMDs were NR). A similar pattern was observed for VAS 100 mm back

pain scores and North American Spine Society pain scores. At 2 years followup, the number and

proportion of participants reporting no leg pain was 37 (82%) among those allocated to revision

endoscopic discectomy and 32 (76%) among those allocated to revision microdiscectomy (P

value for comparison NR).

Functioning/disability

North et al. reported qualitatively that higher levels of impairment from pain in performing

everyday activities among participants allocated to repeat lumbosacral decompression compared

with participants allocated to spinal cord stimulation; however, actual numeric values of

impairment were NR and differences were reported as nonsignificant.47 Ruetten et al. reported

outcomes using the Oswestry disability index.48 Improvements in disability from baseline to 12

weeks, 26 weeks, 52 weeks, and 2 years were observed among participants allocated to revision

endoscopic discectomy and among participants allocated to revision microdiscectomy. The

between-group differences were reported as not significant (AMDs and P values for between-

group differences NR).

Quality of life

Neither study reported outcomes related to overall quality of life.


North et al. did not report any outcomes related to neurologic symptoms. Ruetten et al. reported

mean North American Spine Society Neurology scores at 12 weeks, 26 weeks, 52 weeks, and 2



years.48 Participants allocated to revision endoscopic discectomy had a mean baseline score of 3

and participants allocated to revision microdiscectomy had a mean baseline of 5. Scores in both

groups improved over time; at 2 years the scores were 2.1 and 2.3, respectively. The differences

between groups were reported as not significant (AMDs and P values for between-group

differences NR).

Return to work

North et al. reported no significant differences in return to work, but actual values were not

reported. Ruetten et al. reported a significant difference between groups in the mean duration of

postoperative disability.48 Among participants allocated to revision endoscopic discectomy the

mean was 4 weeks (SD NR) and among participants allocated to revision microdiscectomy the

mean was 7.4 weeks (SD NR) (P < 0.01).


North et al. reported on the frequency and proportion of successful treatment over a mean

follow-up time of 2.9 years (range 1.8 years to 5.7 years).47 Success was defined as at least 50%

pain relief and patient satisfaction with treatment. A significant difference in treatment success

was observed (P < 0.01). Among those allocated to repeat lumbosacral decompression,

successful treatment was observed in 3 (12%). Among those allocated to spinal cord stimulation,

successful treatment was observed in 9 (47%). Ruetten et al. reported on the frequency and

proportion of patient satisfaction with surgery and whether participants would undergo the

operation again. Among those allocated to revision endoscopic discectomy, 43 (95%) were

satisfied; among those allocated to revision microdiscectomy 36 (86%) were satisfied (P value

for comparison NR).

3.3 Safety

Safety Question 1

In adults with symptomatic lumbar radiculopathy, what are the adverse events associated with

surgical interventions?

All 24 RCTs included for EQ1 and the two RCTs included for EQ2 also provided evidence for

safety outcomes. A summary of included studies is provided in Tables 3 and 4. Appendix C,

Tables C-1 and C-2 provide detailed individual study and population characteristics. Appendix

C, Table C-5 provides detailed individual study outcomes related to safety.

3.3.1 Study Characteristics

The study characteristics for the 24 RCTs included for safety outcomes were previously

described in EQ1 and EQ2.

3.3.2 Findings

3.3.2.1 Mortality


Six RCTS reported on mortality.5,21,22,30,31,39 All of these studies were rated as low risk of bias for

this specific outcome. Table 24 summarizes the findings and strength of evidence related to



mortality. A description of findings follows this table. Surgical mortality is not relevant as a

comparative outcome given the non-surgical comparison group. Thus, the strength of evidence

for surgical mortality reflects our certainty about the absolute incidence of surgical mortality in

the surgical intervention group.

Table 24. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for mortality in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Surgical mortality

5 RCTs Not seriousa


No studies reported any surgery-related deaths.21,22,30,31,39

⨁⨁◯◯

LOWc

NAc

All-cause mortality (medium- and long-term)

3 RCTs Not seriousa

Not serious Not serious Very seriousd

All-cause mortality similar between groups. Gerszten et al.39 1 death in each treatment group at 26w, unrelated to surgery. Weinstein et al.[SPORT]21(N=501) 3 deaths (1.91%) vs. 4 deaths (2.63%) by 8y. Weber et al.5(N=126) 3 deaths (5.0%) vs 0 (0%) by 10y.

⨁⨁◯◯

LOW

No difference

a Though all of these trials were rated as either having some concerns for bias or high risk of bias for efficacy outcomes and other

safety outcomes, the risk of bias for mortality outcomes are low since non-comparative surgical mortality and all-cause mortality

outcomes are unlikely to be influenced by lack of participant or outcome assessor blinding or crossovers.

b Optimal information size criterion not met: no events occurred.

c Because the comparator intervention is nonsurgical; this strength of evidence rating reflects the absolute incidence of surgical

mortality, not the relative incidence with respect to a comparator.

d Optimal information size criterion not met: very rare events occurred.

Abbreviations: N = number; RCT = randomized controlled trial; SPORT = Spine Patient Outcomes Research Trial; w =

week(s); y = year(s)

Of the five RCTs that reported surgical mortality, no studies reported any deaths relating to

percutaneous disc decompression,39 microdiscectomy,22,31 discectomy,30 or

discectomy/microdiscectomy procedures.21 Three RCTs reported all-cause mortality.5,21,39

Weinstein et al. [SPORT] reported three deaths (1.91%) among participants allocated to

discectomy/microdiscectomy and four deaths (2.63%) among participants allocated to

conservative management at 8 years.21 Weber et al. reported three deaths (5.0%) among

participants allocated to discectomy and no deaths among participants allocated to conservative

management at 10 years.5 One death was due to cancer and two due to heart disease. Gerszten et

al. reported one death in each intervention group at 26 weeks followup because of myocardial

infarction and acute pyelonephritis.39


Five RCTs reported mortality outcomes.23,27,29,32,41 Table 25 summarizes the findings and

strength of evidence related to mortality for this comparison. A description of findings follows

this table.



Table 25. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for mortality in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Surgical mortality

5 RCTs Not seriousa


No studies reported any surgery-related deaths.23,27,29,32,41

⨁⨁◯◯

LOW

No difference

All-cause mortality

1 RCT Not seriousa

Not seriousc Not serious Very seriousd

Ruetten et al. 29 (N=200) reported one death (0.5% of total) unrelated to treatment but did not specify which surgical group the death occurred.

⨁⨁◯◯

LOW

No difference

a Though these trials were rated as some concerns for bias or high risk for bias for efficacy outcomes and other safety outcomes,

the risk of bias for mortality outcomes is low as these outcomes are unlikely to be influenced by lack of participant or outcome

assessor blinding or crossovers. Some trials remain as some concerns for bias because of inadequate randomization.


c Not applicable as only 1 study.

d Optimal information size criterion not met: very rare events and does not report which treatment group the death occurred in.

Abbreviation: N = number; RCT = randomized controlled trial

No surgery-related deaths were reported in the RCTs that compared percutaneous endoscopic

discectomy,32 endoscopic discectomy,29 microendoscopic discectomy,23,27 or video-assisted

arthroscopic microdiscectomy41 to microdiscectomy or discectomy. One RCT reported all-cause

mortality.29 Ruetten et al. reported one death (0.5%) unrelated to surgery; the authors did not

specify whether this death occurred among participants allocated to the minimally invasive

surgery or among participants allocated to microdiscectomy.


Only one RCT reported a surgical mortality. Teli et al. reported no surgical deaths in either

group.27 No RCTs reported all-cause mortality. Table 26 summarizes the findings and strength of

evidence related to mortality for this comparison.



Table 26. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for mortality in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Surgical mortality

1 RCT Not seriousa

Not seriousb Not serious Very serious c

No surgery-related deaths reported in either treatment group in Teli et al.27(N=142).

⨁⨁◯◯

LOW

No difference

All-cause mortality

0 RCTs

- a Though this trial was rated as some concerns for bias for efficacy outcomes and other safety outcomes, the risk of bias for

mortality outcomes is low as these outcomes are unlikely to be influenced by lack of participant or outcome assessor blinding.

This RCT was rated as some concerns for bias. Sources of bias: lack of participant and outcome assessor blinding.

b Not applicable as only 1 study.

c Optimal information size criterion not met: no events occurred.

Abbreviations: N = number; RCT = randomized controlled trial

D. Revision surgery compared with spinal cord stimulation or an alternative revision surgery

Ruetten et al. compared revision endoscopic discectomy with revision microdiscectomy and

reported no deaths (surgery-related or all-cause mortality)48 North et al. reported no deaths

among participants allocated to repeat lumbosacral decompression and one death (2%) unrelated

to treatment (sudden cardiac event) among participants allocated to spinal cord stimulation at 26

weeks.47 Table 27 summarizes the findings and strength of evidence related to mortality for this

comparison.

Table 27. Summary of findings and strength of evidence ratings of revision surgery for mortality in persons with recurrent lumbar radiculopathy (SQ1)


Summary of Findings CERTAINTY/ Direction of effect

№ of Studies

Risk of Bias


Surgical mortality

2 RCTs Not seriousa


Ruetten et al.48 (N=100) and North et al.47(N=26 allocated to surgery) reported no surgery-related deaths.

⨁⨁◯◯

LOW

No difference

All-cause mortality

2 RCTs Not seriousa

Not serious Not serious Very seriousc

Ruetten et al.48 (N=100) reported no deaths in either surgical group. North et al47 (N=50 total) reported 1 death among spinal cord stimulation participants; the death was because of a sudden cardiac event.

⨁⨁◯◯

LOW

No difference

a Though these trials were rated as high risk of bias for efficacy outcomes and other safety outcomes, the risk of bias for mortality

outcomes is low as these outcomes are unlikely to be influenced by lack of participant or outcome assessor blinding. However,

some concerns for bias remain because of inadequate randomization.48


c Optimal information size criterion not met: only 1 death occurred in 1 study.




3.3.2.2 Surgical Morbidity


All but one5 RCT reported surgical morbidity outcomes.21,22,30,31,35,39 Table 28 summarizes the

findings and strength of evidence related to the absolute incidence of surgical morbidity in the

surgical intervention group as comparative surgical morbidity outcomes are not relevant with a

non-surgical comparison group. A description of findings follows this table.

Table 28. Summary of findings and strength of evidence ratings for surgical morbidity in persons with symptomatic lumbar radiculopathy who undergo surgical intervention (SQ1)



№ of Studies

Risk of Bias


Surgical morbidity

6 RCTs Not seriousa


Surgical complications among participants allocated to surgical groups were generally rare. The largest trial, Weinstein et al.[SPORT]21 (N=245 allocated to surgery), reported 10 (4%) dural tears or spinal leaks, 4 (1.6%) superficial wound infections, 1 (0.4%) vascular injury, 2 (0.8%) other intraoperative complication, and 9 (3.6%) other postoperative complication. Gerszten et al.39(N=90 total) reported 5 (11%) adverse events among participants who underwent surgery and 7 (18%) among participants who underwent epidural steroid injection (calculated P=0.55). Osterman et al.31(N=28 allocated to surgery) reported 1 (3.6%) case of urosepsis. Peul et al.30(N=141 allocated to surgery) reported 2 dural tears and 1 wound hematoma. Erginousakis et al.35(N=31 allocated to surgery) and McMorland et al.22(N=20) reported 0 adverse events.

⨁⨁◯◯

LOWc NAc

a Though all of these trials were rated as either having some concerns for bias or high risk of bias for efficacy and other safety

outcomes, the risk of bias for this outcome is low since non-comparative surgical morbidity outcomes are unlikely to be

influenced by lack of participant or outcome assessor blinding or crossovers.

b Optimal information size criterion not met: events were rare in most studies.

c Because the comparator intervention is nonsurgical; this strength of evidence rating reflects the absolute incidence of adverse

events related to surgery, not the relative incidence with respect to a comparator.

Abbreviations: N = number; RCT = randomized controlled trial; SPORT = Spine Patient Outcomes Research Trial

Surgical complications were generally rare among participants who underwent surgical

intervention. Specific study findings:

Weinstein et al. [SPORT] reported 10 (4.0%) dural tear or spinal fluid leaks, 4 (1.6%)

superficial postoperative wound infection, 1 (0.40%) vascular injury, 2 (0.81%) other

intraoperative complications, and 9 (3.6%) other unspecified postoperative complications

among participants who underwent microdiscectomy.21

Gerszten et al. reported 5 (11%) procedure-related adverse events among participants who

underwent plasma disc decompression participants, compared with 7 (18%) procedure-

related adverse events among participants who underwent epidural steroid injection

participants.39 The authors used a broad definition for adverse events that included pain at



the injection site, increased radicular pain, increased weakness, increased back pain, light

headedness, and muscle tightness or spasms.

Other reported morbidity outcomes among participants allocated to surgical interventions

include one case of urosepsis (3.6%) reported by Osterman et al.31 and one wound

hematoma and two dural tears (combined 1.6%) reported by Peul et al..30

McMorland et al. reported no operative complications among participants who underwent

microdiscectomy22 and Erginousakis et al. reported no operative complications among

participants who underwent percutaneous disc decompression.35


Ten RCTs23,26-29,32,34,37,38,41 reported surgical morbidity outcomes. Because comparative surgical

morbidity outcomes are likely to be influenced by the lack of outcome assessor blinding and

studies did not provide any detail for us to assess whether adverse event ascertainment was

equal, valid, and reliable, we rated this outcome as high risk of bias for most studies. Table 29

summarizes the findings and strength of evidence related to morbidity for this comparison. A


Table 29. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for surgical morbidity in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Surgical morbidity

10 RCTs

Very seriousa


The most commonly reported complications were dural tears and spinal fluid leaks. Between-group differences were generally similar between groups with one exception—.Ruetten et al.29 (N=200) found significantly fewer complications (P<0.05) in participants who underwent endoscopic discectomy compared with microdiscectomy.

⨁◯◯◯


a The risk of bias for this outcome is high since surgical morbidity outcomes are likely to be influenced by lack of outcome

assessor blinding and studies generally did not report enough detail to assess whether adverse event ascertainment was equal,

valid, and reliable in both study arms.

b Optimal information size criterion not me: events are rare, minimally important differences not clear, and most studies probably

do not have sample sizes to detect small to modest differences.


The most common complications reported were those relating to dural tear and spinal fluid leak.

In nine of the 10 RCTs, morbidity incidence was similar between groups, though few reported

statistical significance testing. One RCT reported significantly fewer complications among

participants who underwent endoscopic discectomy compared to participants who underwent

microdiscectomy.29 Specific study findings:

Ruetten et al.29 reported significantly fewer complications (P<0.05) among participants

who underwent endoscopic discectomy compared to participants who underwent

microdiscectomy participants. Complication included transient postoperative dysesthesia



(3.3% vs 5.7%), postoperative bleeding (0% vs 2.3%), delayed wound healing (0% vs

1.1%), and soft tissue infection (0% vs 1.1%).

In Arts et al., dural tears were the most common intraoperative complication reported in

both groups, though the frequency did not differ between groups (actual values NR,

P=0.18) and both intraoperative complications (11% vs. 9%, P=0.27) and postoperative

complications (11% vs. 9%, P=0.47) were not different between groups.38

Brouwer et al. reported 3 cases of transient nerve root injury among participants who

underwent laser disc decompression and 6 cases of adverse events (3 CSF leaks, 1 transient

nerve root injury, and 1 surgery at wrong level) among participant who underwent

microdiscectomy (P values NR).37 In addition, 5 (9%) cases of technical failure were

observed among participants who underwent percutaneous laser disc decompression.37

Franke et al. reported 2 dural tears among participants who underwent microscopically

assisted percutaneous nucleotomy and 3 among participants who underwent

microdiscectomy (calculated P=0.67).34

Ryang et al. also reported dural tears (0 among participants who underwent trocar

microdiscectomy and 2 among those who underwent microdiscectomy, calculated

P=0.49).28

Hermantin et al. reported 1 spinal fluid leak among participants who underwent discectomy

and 0 among participants who underwent video-assisted arthroscopic microdiscectomy; no

infections or neurovascular injuries were reported in either group.41

Huang et al. reported 1 nerve root sleeve tear among participants who underwent

microendoscopic discectomy and 0 cases among participants who underwent discectomy.23

Teli et al. compared microendoscopic discectomy to microdiscectomy and reported dural

tears (6 vs 2), root injury (2 vs 0), spondylodiscitis (1 vs 0), and worsening motor deficit (2

vs 1) and wound infection (0 vs. 4) among participants who underwent microendoscopic

discectomy and microdiscectomy, respectively.27

Thome et al. also reported similar frequency of complications among those who underwent

sequestrectomy compared with those who underwent microdiscectomy (0 vs. 0

intraoperative complications, 0 vs. 1 wound infection, 1 vs.1 nerve root sheath tear, 1 vs. 0

dural leak, 1 vs. 0 discitis).26

Mayer et al. reported no complications in either surgical group.32


Three RCTs reported surgical morbidity,25,27 but one33 did not report by group. Table 30

summarizes the findings and strength of evidence related to morbidity for this comparison. In

one RCT, the overall frequency of surgical infection was 6.3%.41 The other two RCTs reported

similar frequency of complications between groups, but no statistical testing was performed. Teli

et al. reported the frequency of dural tear (2 vs. 2), nerve root injury (0 vs. 0), wound infection (4

vs. 3), and worsening motor deficit (1 vs. 0), among participants who underwent



microdiscectomy compared with participants who underwent discectomy, respectively.27

Tullberg et al. reported the frequency of nerve root sheath (1 vs. 1), dural leak (1 vs. 0), and

discitis (1 vs. 0) among participants who underwent microdiscectomy compared with participants

who underwent discectomy, respectively.25

Table 30. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for surgical morbidity in persons with symptomatic lumbar radiculopathy (SQ1)


Summary of Findings CERTAINTY /Direction of effect

№ of Studies

Risk of Bias


Surgical morbidity



No between-group differences in various surgical morbidity outcomes. Teli et al.27 (N=142) reported 2 dural tears in each group, no differences in nerve root injury, wound infection, spondylodiscitis or worsening motor function (all P values > 0.37 for comparisons). Tullberg et al.25 (N=60) reported 1 nerve root sheath tear in each group, and 1 dural leak and 1 discitis among participants who underwent microdiscectomy and 0 among those who underwent discectomy (calculated P values were NS). Henriksen et al.41 (N=80) reported 5 (6.3%) wound infections overall (NR by group).

⨁◯◯◯


a The risk of bias in these studies for this outcome is high since surgical morbidity outcomes are likely to be influenced by lack of

outcome assessor blinding and studies generally did not report enough detail to assess whether adverse event ascertainment was

equal, valid, and reliable in both study arms.

b Optimal information size criterion not met: rare to no events occurred and most study sample sizes are likely to small to detect

small to modest differences.

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial


Two RCTs that compared revision surgery to alternative interventions reported surgical

morbidity outcomes. Table 31 summarizes the findings and strength of evidence related to

surgical morbidity for this comparison. Ruetten et al. compared revision endoscopic discectomy

to revision microdiscectomy and reported significantly less serious complications (not further

described) among revision endoscopic discectomy participants (6% vs 21%).48 All other

complications, including dural injury (1 vs 3), transient postoperative dysesthesia (2 vs 5),

delayed wound healing (0 vs 2), and soft tissue infection (0 vs 1) occurred with less frequency

among participants who underwent revision endoscopic discectomy. North et al. reported 0 site

infections among repeat lumbosacral decompression participants and 1 (4.2%) among

participants who underwent spinal cord stimulation.47 No other surgical complications were

noted.



Table 31. Summary of findings and strength of evidence ratings of revision surgery for surgical morbidity in persons with recurrent lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Surgical morbidity


Not serious Not serious Seriousb Ruetten et al.48 (N=100) reported significantly less serious complications (not further described) in participants who underwent revision endoscopic discectomy compared to revision microdiscectomy group (6% vs 21%, P<0.05). Other complications reported: dural injury (1 vs. 3), transient postoperative dysesthesia (2 vs. 5), soft tissue infection (0 vs. 1). North et al.47 (N=50) reported 1 (4.2%) site infection among participants who underwent spinal cord stimulation compared to 0 among participants who underwent revision microdiscectomy. No other surgical complications were reported.

⨁◯◯◯

VERY LOW Favors minimally-invasive surgeryc

No differenced

a The risk of bias in these studies for this outcome is high since surgical morbidity outcomes are likely to be influenced by lack of

outcome assessor blinding and studies generally did not report enough detail to assess whether adverse event ascertainment was

equal, valid, and reliable in both study arms.

b Optimal information size criterion may not have been met: rare or no events occurred in 1 study.

c Favors revision endoscopic discectomy over microdiscectomy.

d No difference between repeat lumbosacral decompression and spinal cord stimulation.


3.3.2.3 Reoperations


Five RCTs reported reoperation rates in participants that were allocated to and underwent the

surgical intervention; some studies also reported reoperations among participants who crossed

over from the nonsurgical intervention to surgery.21,22,30,31,35 Table 32 summarizes the findings

and strength of evidence related to reoperation for this comparison. Reoperations is not relevant

as a comparative outcome given the non-surgical comparison group. Thus, the strength of

evidence for reoperation reflects our certainty about the absolute incidence of reoperations

among those who underwent surgery, whether initially allocated to the surgical group or among

those who crossed over to surgery at some point during the trial. A description of findings

follows this table.



Table 32. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for reoperations in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Reoperations



The incidence of reoperations in study groups varied between 0% to 10.1%. Weinstein et al.[SPORT]21 (N=245 plus crossovers) 25 (10.1%) had a reoperation within 2y. Peul et al.30 (N=142 allocated to surgery) 7 (6%) had reoperation by 2y and 9 (7%) by 5y. Osterman et al.31 (N=28) allocated to surgery) 2 (7.1%) reoperations and Erginousakis35 (N=31 allocated to surgery) 1 (3.2%). McMorland et al.22 (N=20 allocated to surgery) reported no reoperations.

⨁◯◯◯

VERY LOWc NAc

a This outcome was rated as high risk of bias in all studies because it includes reoperations among those who crossed over.

b Optimal information size criterion not met: unclear what a minimally important difference is, all but the largest trial have

sample sizes that are unlikely to detect small to modest differences.

c Because the comparator intervention is nonsurgical; this strength of evidence rating reflects the absolute incidence of

reoperations related to surgery, not the relative incidence with respect to a comparator.

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial; SPORT = Spine Patient Outcomes Research

Trial; y = year(s).

The incidence of reoperations across the five RCTs varied from 0% to 10.1%. Specific study

findings:

Peul et al. reported that 7 (6%) participants allocated to microdiscectomy had reoperations

for recurrent sciatic within 2 years and 9 (7%) by 5 years.30 Among participants allocated

to conservative management who crossed over to receive surgery, 4 (6%) underwent a

reoperation by 2 years and 8 (12%) by 5 years.

Weinstein et al. [SPORT] reported that 25 (10.1%) of participants who underwent

microdiscectomy/discectomy (including crossover) had reoperations for recurrent

herniation, complication, or other reasons within 2 years.21

Erginousakis et al. reported 1 reoperation among participants allocated to percutaneous

disc decompression but the follow-up time period was not reported.35

McMorland et al. reported no reoperations among participants allocated to

microdiscectomy.22 Osterman et al. reported 2 (7.1%) reoperations among participants who

underwent microdiscectomy.


Ten RCTs reported reoperation rates.26-29,32,34,36-38,41 Table 33 summarizes the findings and

strength of evidence related to reoperation for this comparison. A description of findings follows

this table.



Table 33. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for reoperations in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Reoperations

10 RCTs

Very seriousa

Seriousb Not serious Very Seriousc

The proportion of participants in each study group who had a reoperation varied from 2% to 64.5%. Arts et al.38 (N=328) reported 23 (15%) among participants allocated to tubular discectomy and 14 (10%) among participants allocated to microdiscectomy at 2y (P=0.22). Between-group differences in other studies varied, but no statistical testing conducted. Most studies reported reasonably similar frequency of reoperations with the exception of Chatterjee et al.36 (N=71) 20 (64.5%) reoperations among participants who underwent automated percutaneous lumbar discectomy vs. 1 (2.5%) reoperation in microdiscectomy group; and Brouwer et al.37 (N=115) 24 (44%) reoperations among participants who underwent percutaneous laser disc decompression compared with 9 (16%) in microdiscectomy (P=0.002).

⨁◯◯◯

VERY LOW No difference/ Favors standard surgery

a One trial was rated as low risk for bias,38 six were rated as some concerns for bias26,27,34,36,37,41and three were rated as high risk

for bias.28,29,32 Sources of bias: lack of participant and outcome assessor blinding in all but the low risk of bias trial, inadequate

randomization and allocation concealment28,29,32 and deviations from intended intervention.32

b The magnitude of difference in two RCTs36,37 was much larger than any of the other 8 RCTs in this body of evidence.

c Optimal information size criterion not met: reoperations rare in some studies, most study sample sizes unable to detect small to

modest differences between groups.


The proportion of participants that had reoperations varied extensively across study groups (from

2.5 % to 64.5). Two studies observed significantly higher frequency of reoperations among

participants who underwent minimally-invasive surgery compared to standard surgery.36,37 These

findings were inconsistent with findings from the other 8 RCTs, which observed similar

incidence of reoperations between surgical groups.

Brouwer et al. reported 24 (44%) reoperations within 52 weeks among those who

underwent percutaneous laser decompression compared with 9 (16%) reoperations among

those who underwent microdiscectomy (calculated P=0.002).37

Chatterjee et al. reported 20 (64.5%) reoperations among participants who underwent

automated percutaneous discectomy and 1 (2.5%) among participants who underwent

microdiscectomy (calculated P<0.001).36

Arts et al. reported 23 (15%) reoperations within 2 years among participants who

underwent tubular discectomy and 14 (10%) reoperations among participants who

underwent microdiscectomy (P=0.22).38

Franke et al. reported 2 (3.9%) reoperations among those who underwent microscopically-

assisted percutaneous nucleotomy (1 was for relapse same level/same side and 1 was for



progressive disc degeneration and segmental instability) and 5 (10.4%) reoperations among

those who underwent microdiscectomy (4 were for relapse and 1 was for progression).34

This difference in proportion was not significant (calculated P=0.26).

Mayer et al. reported 3 (15%) reoperations among those who underwent percutaneous

endoscopic discectomy compared with 1 (5%) reoperations among those who underwent

microdiscectomy (calculated P=0.61).32

Ruetten et al. reported 7 (7.7%) reoperations among participants who underwent

endoscopic discectomy compared with 10 (11.5%) reoperations among those who

underwent microdiscectomy (calculated P=0.45).29

Ryang et al. reported 2 (6.6%) reoperations among participants who underwent trocar

discectomy; one was during the initial hospital stay and one was within 6 weeks. Among

participants who underwent microdiscectomy, 4 (13.3%) reoperations were performed (one

during the initial hospital stay and one at 8 weeks, 28 weeks, and 1.2 years.28 The

difference in proportion of participants who underwent reoperations between groups was

not significant (calculated P=0.67).

Teli et al. reported 8 (11.4%) reoperations among participants who underwent

microendoscopic discectomy compared with 3 (4.2%) reoperations among participants who

underwent microdiscectomy (calculated P=0.13).27

Thome et al. reported 2 (5%) reoperations for recurrent herniation within 1.5 years among

participants who underwent sequestrectomy compared with 4 (10%) reoperations among

participants who underwent microdiscectomy.26 This difference was not significant

(calculated P=0.68). Reherniations (with or without reoperation) were reported in 12.5% of

participants allocated to sequestrectomy compared with 10.5% of participants allocated to

microdiscectomy (calculated P=1.0).

Hermantin reported 1 (3.3%) reoperations among participants who underwent video-

assisted arthroscopic microdiscectomy for treatment of mild lateral stenosis that had not

been recognized at the time of surgery. Among participants who underwent discectomy, 2

(6.7%) reoperations were performed; one was for repair of a dural sac/spinal fluid leak and

the other was for persistent radicular symptoms.41 The difference in proportion between

groups was not significant (P=1.0).


Two RCTs reported on reoperations.25,27Table 34 summarizes the findings and strength of

evidence related to reoperation for this comparison. Tullberg et al. reported 1 (3.3%) reoperation

by 52 weeks in each surgical group (microdiscectomy and discectomy).25 Teli et al. reported 3

(4.2%) reoperations among participants who underwent microdiscectomy compared with 2 (3%)

among participants who underwent discectomy (calculated P=1.0).27



Table 34. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for reoperations in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Reoperations

2 RCTs Seriousa Not serious Not serious Very seriousb

No between-group differences in frequency of reoperations. Tullberg et al.25 (N=60) reported 1 (3.3%) reoperation in each surgical group by 52w (calculated P=1.0). Teli et al.27 (N=142) reported 3 (4.2%) reoperations among participants who underwent microdiscectomy compared with 2 (3%) among participants who underwent discectomy (calculated P=1.0).

⨁◯◯◯


a Both studies were rated as some concerns for risk of bias. Sources of bias: lack of participant and outcome assessor blinding.

b Optimal information size criterion not met: reoperations were rare events, study sample sizes unable to detect small to modest

differences between groups.



Two RCTs reported reoperations and reported similar frequencies between surgical groups.

Table 35 summarizes the findings and strength of evidence related to reoperation for this

comparison. North et al. reported 0 reoperations among participants who underwent repeat

lumbosacral decompression and 3 (12.5%) hardware revisions among participants who

underwent spinal cord stimulation.47 Ruetten et al. reported 2 reoperations at 2 years among

participants who underwent revision endoscopic discectomy compared with 3 reoperations

among participants who underwent revision microdiscectomy.48

Table 35. Summary of findings and strength of evidence ratings of revision surgery for reoperations in persons with recurrent lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias


Reoperations



Ruetten et al.48 (N=100) reported 2 (4%) reoperations among participants who underwent revision endoscopic discectomy and 3 (6%) reoperations among participants who underwent revision microdiscectomy (calculated P=0.67). North et al.47 (N=50) reported 0 reoperations among participants who underwent repeat decompression, and 3 (12.5%) hardware revisions among participants who underwent spinal cord stimulation (P value NR).

⨁◯◯◯


a Both trials were rated as high risk of bias. Sources of bias: lack of participant and outcome assessor blinding and extensive

crossovers and differential attrition47 and inadequate randomization and allocation concealment.48

b Rare events, optimal information size criteria




3.3.2.4 Persistent Opioid Use


Only one RCT reported outcomes related to persistent opioid use.39 Table 36 summarizes the

findings and strength of evidence related to persistent opioid use for this comparison. Gerszten et

al. reported that reduction in use of narcotics was not significantly different between participants

who underwent percutaneous disc decompression and those who underwent conservative

management participants at 26 weeks (actual values NR, P value NR).

Table 36. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for persistent opioid use in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias



1 RCT Very seriousa


Gerszten et al.39 (N=90) reported no significant difference in reduction in narcotics between participants who underwent percutaneous disc decompression and compared with conservative management participants at 26w (actual values NR, P value NR).

⨁◯◯◯


a. Risk of bias is high for 12 week and later outcomes. Sources of bias: lack of participant and outcome assessor blinding, and

deviations from intended interventions.

b. Not applicable as only 1 study.

c Optimal information size criterion not met: actual values and measures of variance not reported.

Abbreviations: N = number; NR = not reported; RCT = randomized controlled trial; w = week(s).


Only one RCT reported outcomes related to persistent opioid use.41 Table 37 summarizes the

findings and strength of evidence related to persistent opioid use for this comparison. The

duration of postoperative narcotic use ranged from 0.43 to 2 weeks (average 1 week) for

participants who underwent video-assisted arthroscopic microdiscectomy and 1 to 8 weeks

(average 3.65 weeks) for participants who underwent discectomy (P value NR).

Table 37. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for persistent opioid use in persons with symptomatic lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias




Hermantin et al.41 (N=60) reported mean duration of postoperative narcotic use in video-assisted arthroscopic microdiscectomy participants of 1 week (range 0.43w to 2w) compared to 3.65 weeks (range 1w to 8w) in participants who underwent discectomy (P value NR).

⨁◯◯◯

VERY LOW Favors minimally-invasive surgery

a. This RCT was rated as some concerns for bias. Sources of bias: lack of participant and outcome assessor blinding.

b. Not applicable as only 1 study.



c. Optimal information size criterion not met: measures of variance or confidence intervals not provided.



No RCTs reported persistent opioid use outcomes.


Two RCTs reported outcomes related to opioid use.47,48 Table 38 summarizes the findings and

strength of evidence related to persistent opioid use for this comparison. Ruetten et al. reported

that postoperative pain medication was significantly reduced with revision endoscopic

discectomy compared to revision microdiscectomy (actual values NR, P<0.01).48 North et al.

reported that by 2.9 years, 15 (58%) participants who underwent repeat lumbosacral

decompression participants reported stable or decreased opioid use compared to 30 (87%)

participants who underwent spinal cord stimulation participants (P=0.025).47

Table 38. Summary of findings and strength of evidence ratings of revision surgery for persistent opioid use in persons with recurrent lumbar radiculopathy (SQ1)



№ of Studies

Risk of Bias




Not seriousb Not serious Not serious Ruetten et al.48 (N=100) opioid use was significantly reduced in participants who underwent revision endoscopic discectomy compared to participants that underwent revision microdiscectomy (P<0.01). North et al.47 (N=50) 15 (58%) participants with stable or decreased opioid use among participants allocated to repeat surgery compared to 30 (80%) participants allocated to spinal cord stimulation at long-term followup (P=0.025).

⨁⨁◯◯ LOW Favors minimally-invasive surgeryc

Favors spinal cord stimulationd

a Both trials were rated as high risk of bias. Sources of bias: lack of participant and outcome assessor blinding and extensive

crossovers and differential attrition47 and inadequate randomization and allocation concealment.48

b Results may differ because of different comparator groups.

c Favors revision endoscopic microdiscectomy over microdiscectomy.

d Favors spinal cord stimulation over repeat lumbosacral decompression.


3.4 Cost and Cost-Effectiveness

Cost Question 1

In adults with symptomatic lumbar radiculopathy, what is the cost-effectiveness of surgical

interventions?

We identified seven eligible studies reporting cost.27,42,49-53 Teli et al., a trial we also included for

efficacy and safety outcomes, reported on surgical costs of three alternative surgical

interventions.27 Five studies reported cost-effectiveness analyses related to RCTs that we also

included for efficacy and safety outcomes.49-53 Lastly, Malter et al. reported a cost-effectiveness



analysis using cost and effectiveness inputs from a variety of sources.42 A summary of included

studies is provided in Table 39. Appendix C, Tables C-1 and C-2 provide detailed individual

study and population characteristics for the cost studies related to included RCTs and Table C-6

provides detailed information related to methods and findings specific to the seven cost studies.

These studies reported cost findings using different currency and base years; thus, we converted

all figures to 2010 U.S. dollars (see Appendix B, Table B-1 for details on conversion) for this

report.

Table 39. Study characteristics of the seven studies that evaluated cost effectiveness of surgery for lumbar radiculopathy (CQ1)

Author (Year); Related RCT

Surgical Intervention (N randomized)

Comparator (N randomized) Key Analysis Parameters Outcomes Reported

Cost-effectiveness of surgery compared with nonsurgical interventions

Van den Hout (2008)49 Related RCT: Peul (2007)30

Discectomy (141)

Conservative management (142)

Year/Currency: 2008 € Discount rate: 0% Time Horizon: 52w Costs included: direct and indirect QOL measure: EQ-5D with U.K. norms

Mean QALY

Mean total costs

Cost/QALY gained

Mean health care costs

Health care costs/QALY gained

Tosteson (2008)50 Related RCT: Weinstein (2006)21 [SPORT]

Discectomy /microdiscectomy (245)

Conservative management (256)

Year/ currency: 2004 U.S.$ Discount rate: 3% Time horizon: 2y Costs included: direct and indirect QOL measure: EQ-5D with U.S. norms Other: Based on pooled data from SPORT RCT and observational cohort.

Mean QALY

Mean total costs

Cost/QALY gained

Mean direct medical costs

Direct medical costs/QALY gained

Malter (1996)42 Discectomy (NA) Nonsurgical management (NA)

Year/currency: 1993 U.S.$ Discount rate: 5% Time horizon: 10y Costs included: Direct medical costs QOL measure: Author developed time-trade off utility measure Other: Efficacy estimates based on an RCT comparing surgery with nonsurgical treatment (Weber et al. (1983)5), and an RCT comparing surgery with chemonucleolysis (Javid et al. (1998)106) and a cohort study comparing surgery with nonsurgical treatment (Atlas et al. (1993)107).

QALY

Costs

Costs/QALY gained

Cost-effectiveness of alternative surgical interventions

Van den Akker (2011)51 Related RCT: Arts (2009)38

Tubular discectomy (167)

Microdiscectomy (161)

Year/Currency: 2008 U.S.$ Discount rate:0% Time horizon: 52w QOL measure: EQ5-D Costs included: direct and indirect

Mean QALY

Mean total costs

Cost/QALY gained

Mean health care costs


(continued)



Table 39. Study characteristics of the seven studies that evaluated cost effectiveness of surgery for lumbar radiculopathy (CQ1) (continued)

Author (Year); Related RCT

Surgical Intervention (N randomized)

Comparator (N randomized) Key Analysis Parameters Outcomes Reported

Van den Akker (2017)52 Related RCT: Brouwer (2015)37

Percutaneous laser disc decompression (57)

Discectomy, with laminotomy as needed (58)

Year/ Currency: 2010 € Discount rate: 0% Time Horizon: 52w QOL measure: EQ-5D with U.S. norms Costs Included: direct and indirect

Mean QALY

Mean total costs

Costs/QALY gained


Stevenson (1995)53 Related RCT: Chatterjee (1995)36

Automated percutaneous lumbar discectomy (31)

Microdiscectomy (40)

Year/Currency: 1992 £ Discount rate: NR Time Horizon: 26w Costs Included: direct and indirect QOL measure: NR Other: effectiveness was assessed on a 4-pt Likert scale by two clinicians (4=excellent, 1=poor). "Successful outcome" was defined as a 3 or 4.

Mean total cost

Cost per successful outcome

Cost per point gained on 4-pt Likert scale of effectiveness

Teli (2010)27 Microendoscopic discectomy (70)

Microdiscectomy (72) Open discectomy (70)

Year/currency: Euros, Year NR Discount rate: NR Time horizon: NA, procedure costs only Costs included: procedure costs only

Mean surgical costs

Abbreviations: N = number; NA = not applicable; NR = not reported; QALY = quality-adjusted life year; QOL = quality of life;

RCT = randomized controlled trial; SPORT = Spine Patient Outcomes Research Trial; U.K. = United Kingdom; U.S. = United

States; w = week(s); y = year(s)

3.4.1 Study Characteristics

One study was a decision analysis based on published cost and effectiveness data.42 One study

provided cost outcomes as part of the main RCT publication.27 Five studies were separately

published cost-effectiveness analyses of RCTs that we included for efficacy and safety

outcomes27,49,51-53 We rated a related RCT as low risk of bias,38 three RCTs as some concerns for

bias,27,36,37 and two RCTs as high risk of bias.21,30 One of these cost analyses, Tosteson et al.,50

used pooled data from the SPORT trial RCT21 combined with the SPORT observational study

(combined N=1,191). Crossovers in the RCT were included in the treatment group to which they

crossed over. Two studies were conducted in the United States;42,50 the rest were conducted in

the Netherlands,49,51,52 Italy,27 and the United Kingdom.53 The time horizon used in studies

ranged from 26 weeks to 10 years.

Three studies provided evidence for the cost-effectiveness of surgery compared with nonsurgical

treatment.42,49,50 Four studies27,51-53 provided evidence for the comparative cost-effectiveness of

alternative surgical interventions, including percutaneous laser discectomy,52 tubular

discectomy,51 and automated percutaneous discectomy53 compared to microdiscectomy and a

three-arm study comparing microendoscopic discectomy, microdiscectomy, and discectomy.27

Four studies calculated mean quality adjusted life years (QALY) using the EQ-5D measure of

health-related quality of life, and one used an author developed time-tradeoff utility measure.42

These five studies included both direct and indirect costs. The other two studies did not measure

QALYs and only reported direct medical or procedure costs27,53



3.4.2 Findings


Three studies reported cost and cost-effectiveness results.42,49,50 Table 40 summarizes the

findings and strength of evidence related to cost-effectiveness for this comparison. A detailed

description of findings follows this table. Two reported findings from both a societal perspective

(direct and indirect costs)49,50 and a payor perspective (direct medical costs); whereas one

reported only from a payor perspective.42 A detailed description of findings follows this table.

Table 40. Summary of findings and strength of evidence ratings comparing surgery to nonsurgical interventions for cost and cost-effectiveness in persons with symptomatic lumbar radiculopathy (CQ1)


Summary of Findingsa CERTAINTY/ № of Studies

Risk of Bias


Cost-effectiveness

3 studiesb

Very seriousc

Not serious Not serious Very seriousd

All studies reported higher QALYs but similar or higher costs among surgical interventions compared to conservative management. The mean cost per QALY gained from the payor perspective ranged from $51,156 to $83,322. Tosteson et al.50 [SPORT](N=1,191) reported total cost per QALY gained at 2y $80,115 (95% CI, $56,167 to $109,662) and direct medical costs per QALY gained $83,322 (95% CI, $65,189 to $106,655). Van den Hout et al.49 (N=283) did not report an ICER for total costs per QALY gained (calculated ICER $-419/QALY gained interpreted as surgery more effective and costs less), the health care costs per QALY gained at 52w was $63,011 (95% CI, $21,516 to $660,847). Malter et al.42 reported undiscounted cost per QALY gained $44,064 (95% CI NR, discounted $51,156 per QALY gained).d

⨁◯◯◯

VERY LOW

a All costs are reported here in 2010 US Dollars. The costs for the year and currency reported in the published studies is in

Appendix C, Table C-6.

b Two studies were cost-effective analyses were conducted concurrent to RCTs; one study was a cost-effectiveness analysis

conducted using effectiveness inputs from published RCTs and cost inputs from a commercial database.

c The RCTs related to two of the cost studies were rated as high risk for bias. Sources of bias: lack of participant and outcome

assessor blinding and extensive crossovers. The cost analyses associated with these studies were rated as some concerns for bias.

d One study did not provide confidence intervals around incremental cost-effectiveness ratio estimate and we were unable to

calculate it based on available data42; the other 2 studies have confidence intervals around their incremental cost-effectiveness

ratio estimate that span the range of cost savings, probably cost-effective, and not at all cost-effective.

Abbreviations: ICER = incremental cost-effectiveness ratio; N = number; QALY = quality-adjusted life year; RCT =

randomized controlled trial; SPORT = Spine Patient Outcomes Research Trial; w = week(s); y = year(s)

All studies reported higher QALYs among participants allocated to surgical interventions, but

similar or higher costs. The mean cost per QALY gained from the payor perspective ranged from

$51,156 to $83,322.

Tosteson et al.50 used cost and effectiveness inputs from Weinstein et al. [SPORT],21 an

RCT conducted in the United States, and reported outcomes at 2 years using 2004 U.S.

dollars ($). The difference in QALYs was 0.21 (95% CI 0.16 to 25), favoring

microdiscectomy/discectomy compared with conservative management. The total cost



among participants who received surgery was $31,561 (95% CI, $29,877 to $33,244) and

the total cost among participants who received conservative management was $15,162

(95% CI, $12,979 to $17,202); the AMD was not reported but we calculated it to be

$16,399 (95% CI, 95% CI, $16,289 to $16,509). The study reported the cost per QALY

gained (also known as the incremental cost-effectiveness ratio, ICER) was $80,115 (95%

CI, $57,167 to $109,662). When limited to direct medical costs, the cost among

participants who received surgery was $23,361 (95% CI, $22,295 to $24,426) and the cost

among participants who receive conservative management was $6,700 (95% CI, $5,355 to

$8,045); we calculated this AMD to be $16,661 (95% CI, $16,590 to $16,732). The direct

medical costs per QALY gained was $83,322 (95% CI, $65,189 to $106,655), which means

that although surgery is more effective, it also costs more. Whether this estimate is cost-

effective depends on the threshold of additional costs that payors are willing to pay for an

additional QALY.

Van den Hout et al.49 used cost and effectiveness inputs from an RCT conducted in The

Netherlands30 and reported outcomes at 52 weeks using 2008 Euros (€). Microdiscectomy

resulted in a higher mean QALY (AMD 0.044 [95% CI, 0.005 to 0.083] but no significant

differences in total costs (AMD -$18.44 [95% CI, $-6,192 to $6,157) compared with

conservative management. The authors concluded that microdiscectomy dominates

conservative management (i.e., is more effective and costs less). However, no cost per

QALY gained (i.e., ICER) was reported for the societal perspective. We calculated the

ICER as $-419 per QALY gained (95% CI unable to be calculated). Using the 95%

confidence intervals provided for the difference in QALYs and the difference in costs by

the published study, we calculated the possible range of this estimate to be from $-

74,602/QALY (best case) to $1,231,400 (worse case). When limited to health care costs

only, the mean cost among participants in the surgical group was $8,646 (SD $5,955) and

$5,851 (SD $6,512) for an AMD of $2,796 (95% CI, $1,294 to $4,288). This results in an

ICER of $63,011 (95% CI, $21,516 to $660,847), which means that surgery was more

effective but also costs more.

Lastly, Malter et al.42 conducted a cost-effectiveness analysis over a 10-year horizon using

effectiveness data inputs from several published RCTs and observational studies of surgery

compared with nonsurgical interventions, including conservative management and

chemonucleolysis comparators, and an author-developed time-tradeoff utility

measure.5,106,107 Health care cost inputs were obtained from a commercial database of U.S.

costs from 1987 to 1989 and were adjusted to 1993 dollars for reporting in the analysis.

The difference in undiscounted QALYs was 0.43 (95% CI NR). The health care costs

associated with surgery were estimated to be $25,684 (95% CI NR) and the costs

associated with nonsurgical management were estimated to be $6,745 (95% CI NR) (AMD

$18,938 [95% CI NR]). The undiscounted cost per QALY gained (ICER) was $44,064

(95% CI NR) and the discounted cost per QALY gained was $51,156 (95% CI NR), which

means that surgery was more effective but also costs more.


Four studies reported cost and cost-effectiveness results.27,51-53 One reported findings only from a

societal perspective,53 one reported findings only from a payor perspective,27 and two reported

findings from both a societal perspective and a payor perspective.51,52 Table 41 summarizes the



findings and strength of evidence related to cost-effectiveness for this comparison. A detailed


Table 41. Summary of findings and strength of evidence ratings comparing minimally invasive surgery to standard surgery for cost and cost-effectiveness in persons with symptomatic lumbar radiculopathy (CQ1)


Summary of Findingsa CERTAINTY/ № of Studies

Risk of Bias


Cost and cost-effectiveness

4 studiesb

Seriousc Seriousd Not serious Very seriouse

Inconsistent findings across studies. Teli et al.27 (N=142) calculated AMD for surgical costs $722 (95% CI, $551 to $892) comparing microendoscopic discectomy to microdiscectomy. Stevens et al.53 (N=70) calculated AMD cost per successful outcome at 26w $3,573 comparing automated percutaneous lumbar discectomy to microdiscectomy. Van den Akker51 (N=325) no significant differences in QALYs, total costs, or health care costs; costs and health care costs per QALY are NR but point estimates for differences in QALYs and costs suggest minimally invasive surgery is less effective and costs more. Van den Akker52 (N=115) no significant differences in QALYs or total costs but some difference in health care costs (AMD $-2,393 (95% CI, $-4,376 to $-409); costs and health care costs per QALY NR, but point estimates for differences in QALYs and costs suggest that minimally invasive surgery may be less effective but also costs less.

⨁◯◯◯

VERY LOW

a All costs are reported here in 2010 US Dollars. The costs for the year and currency reported in the published studies is in

Appendix C, Table C-6.

b All studies were conducted concurrent to RCTs.

c The risk of bias in one of the related RCTs was low and the risk of bias in the other three related RCTs was some concerns.

Sources of bias: lack of participant and outcome assessor blinding. The risk of bias specific to the cost analyses was some

concerns.

d Three studies show that minimally invasive surgery is slightly more expensive, one study shows that minimally invasive surgery

is significantly less expensive. Inconsistency in findings related to cost-effectiveness among the two studies that reported

QALYs.

e Confidence intervals span thresholds of cost effectiveness and cost savings, some studies do not report estimates for the

incremental cost-effectiveness ratio or measures of variance.

Abbreviations: ICER = incremental cost-effectiveness ratio; N = number; NR = not reported; QALY = quality-adjusted life

year; RCT = randomized controlled trial; w = week(s); y = year(s)

The findings were inconsistent across studies.

Teli et al. reported surgical costs using Euros (year unspecified) as part of reporting results

from an RCT conducted in Italy. The mean surgical cost among participants who

underwent microendoscopic discectomy was $3,878 (SD $580) and was $3,156 (SD $438)

among participants who underwent microdiscectomy. The AMD was not reported but we

calculated it to be $722 (95% CI, $551 to $892), which suggests that the minimally

invasive approach was more expensive than microdiscectomy.



Stevenson et al.53 used cost inputs from an RCT conducted in the United Kingdom36 and

reported outcomes at 26 weeks using 1992 British pounds (£). This study reported costs

and cost per successful outcome, where two clinicians measured success using a 4-point

Likert scale (1=poor, 4=excellent) (i.e., MacNab criteria). Success was defined as a 3 or 4

on this scale. In the related RCT, 9 (29%) of participants allocated to automated

percutaneous lumbar discectomy had a successful outcome compared with 32 (80%) of

participants allocated to microdiscectomy (p< 0.001).36 The mean differences in success

scores were not reported in the main RCT or in the cost study. The mean total cost was

$6,340 among participants allocated to automated percutaneous lumbar discectomy and

$4,288 among participants allocated to microdiscectomy (AMD NR, calculated to be

$2,052). Cost estimates for both groups included the cost of additional surgeries for failed

initial surgery. The cost per successful outcome among participants allocated to automated

percutaneous lumbar discectomy was $8,931 versus $5,358 among participants allocated to

microdiscectomy (calculated AMD $3,573). The authors calculated the cost per additional

point gained on the 4-pt. Likert scale of success as $3,770 vs. $2,091 for the two surgical

groups, respectively (calculated AMD $1,688).

Van den Akker et al.51 used cost and effectiveness inputs from an RCT conducted in The

Netherlands38 and reported outcomes at 52 weeks using 2008 U.S.$. The mean difference

in QALYs was not significantly different between participants who underwent tubular

discectomy compared with participants who underwent microdiscectomy (AMD -0.12

[95% CI, -0.046 to 0.021]). Similarly, no significant difference in mean total costs or health

care costs were observed (AMD $1,510 [95% CI, -$1,352 to $4,373] and AMD $466

([95% CI, -$246 to $1,178], respectively). The total costs per QALY (ICER) was not

reported. Point estimates for the difference in effectiveness and cost were not significant

but the direction of their differences suggests that microdiscectomy dominates (i.e., is more

effective and costs less) minimally invasive surgery.

Van den Akker et al.52 also reported cost and cost-effectiveness from a different RCT

conducted in The Netherlands.37 This analysis reported outcomes at 52 weeks using 2010

Euros. The mean difference in QALYs was -0.033 (95% CI NR, p=0.27) comparing

percutaneous laser disc discectomy with microdiscectomy. The difference in total costs was

$-3,215 (95% CI, -$10,294 to $3,865) and the difference in health care costs was $-2,393

(95% CI, -$4,376 to $-409). The total costs per QALY (ICER) and health care costs per

QALY were not reported. Point estimates for the difference in effectiveness and total costs

were not significant but the direction of their differences suggests that microdiscectomy

dominates (i.e., is more effective and costs less) minimally invasive surgery.


Teli et al. also reported surgical costs of microdiscectomy to discectomy, but did not report cost-

effectiveness.27 The cost of microdiscectomy was $3,156 (SD $438) and the cost of discectomy

was $2,976 (SD $322). The AMD was not reported but we calculated it to be $65 (95% CI, $52

to $307). Table 42 summarizes the findings and strength of evidence related to cost for this

comparison.



Table 42. Summary of findings and strength of evidence ratings comparing microdiscectomy to discectomy for direct surgical costs in persons with symptomatic lumbar radiculopathy (CQ1)


Summary of Findings CERTAINTY № of Studies

Risk of Bias


Direct surgical costs

1 studya Seriousb Not seriousc Not serious Not serious Teli et al.27 (N=142) reported slightly higher costs for microdiscectomy compared to discectomy ($3,156 (SD $438) vs $2,976 (SD $322). Calculated AMD $65 (95% CI, $52 to $307).

⨁⨁⨁◯

MODERATE

a Cost effectiveness analysis conducted concurrent to an RCT

b The related RCT was rated as some concerns for bias Sources of bias: lack of participant and outcome assessor blinding. The

cost analysis associated with this RCT was rated as some concerns for bias.

c Not applicable as only 1 study.

Abbreviations: AMD = absolute mean difference; N = number; RCT = randomized controlled trial; SD = standard deviation;

3.5 Clinical Practice Guideline Synthesis

We identified 14 relevant clinical practice guidelines (CPGs) or interventional procedures

guidance related to the use of surgical interventions for lumbar radiculopathy; these are

summarized in Table 43. We identified CPGs from the American Pain Society (2009), the

American Society of Interventional Pain Physicians (2013), the North American Spine Society

(2012), the National Institute for Health and Care Excellence (U.K.) (2016), and the American

College of Occupational and Environmental Medicine (2016). In addition, we identified nine

“interventional procedures guidance” related to minimally invasive spine procedures from the

National Institute for Health and Care Excellence (U.K.). The CPGs vary with respect to how

they were developed and the types of studies that were included to inform clinical

recommendations. For example, many of the interventional procedure guidance documents were

partly based on case series, and most included cohort studies or other nonrandomized trial study

designs. The strength of evidence that the guidelines are based on also varied. Only one CPG

(National Institute for Health and Care Excellence, “Low Back pain and Sciatica in over 16s”)

used the GRADE approach; this CPG rated the evidence as low or very low for nearly all

comparisons and outcomes considered.46

Overall, the guidelines we identified were in general agreement about considering discectomy or

microdiscectomy (and related decompressive procedures) as acceptable treatment based on

evidence that it improves outcomes in the short to medium term. One guideline specifies that this

surgery can be considered when symptoms have not improved with conservative therapy.46

Another guideline suggests that conservative therapy is reasonable for patients with

nonprogressive symptoms who wish to delay surgery.54 The guideline recommendations relating

to minimally invasive spine surgery varied; one did not consider these specific procedures within

their scope.46 Three of the guidelines were developed 5 or more years ago; thus may not include

the most recent evidence for these procedures.55-57



Table 43. Clinical practice guidelines related to lumbar radiculopathy or herniated intervertebral lumbar disc

Organization Guideline Title (Year) Guideline Qualitya Recommendationb Evidence Base

Rating/Strength of Evidence Narrative Assessment

American Pain Society Interventional Therapies, Surgery, and Interdisciplinary Rehabilitation for Low Back Pain (2009)55 Quality Rating: 5 out of 7

Open discectomy or microdiscectomy for radiculopathy with prolapsed disc.

4 RCTs comparing surgery to conservative management

Level B/Goodc Moderate net benefit for short-term outcomes (up to 12w) only

Insufficient evidence for determining superiority of open vs. micro approaches.

Insufficient evidence to evaluate alternative surgical methods, including laser- or endoscopic-assisted techniques.”

National Institute for Health and Care Excellence (United Kingdom) Low back pain and sciatica in over 16s: assessment and management-Invasive treatments (2016)46 Quality Rating: 6 out of 7

Consider spinal decompression for sciatica (includes laminectomy, foraminotomy, and/or discectomy) when nonsurgical treatment has not improved pain or function and their radiological findings are consistent with sciatica symptoms.

9 RCTs comparing surgery to nonsurgical treatment including epidural steroids, analgesics and anti-inflammatory medication, physical therapyd 4 cohort studies comparing decompression to fusion or conservative treatment

Low or very low for nearly all comparisons and outcomese Sciatic symptoms tend to improve naturally with time without treatment, but earlier symptom resolution with surgical intervention should be an option for people.

American Society of Interventional Pain Physicians An Update of Comprehensive Evidence-Based Guidelines for Interventional Techniques in Chronic Spinal Pain (2013)20,56 Quality Rating: 4 out of 7

For lumbar disc prolapse, protrusion, and extrusion: automated percutaneous lumbar decompression (APLD), percutaneous lumbar disc decompression (PLDD), and mechanical decompression with nucleoplasty are recommended in select cases.

19 observational studies for APLD. 15 observational studies for laser-assisted PLDD 1 SR of 3 observational studies PLDD with DeKompressor. 1 RCT and 14 observational studies for nucleoplasty.

The evidence is limited for APLD, PLLDD, and percutaneous disc decompression with DeKompressor. The evidence is limited to fair for mechanical lumbar disc decompression with nucleoplasty.

(continued)



Table 43. Clinical practice guidelines related to lumbar radiculopathy or herniated intervertebral lumbar disc (continued)



North American Spine Society Clinical Guidelines for Diagnosis and Treatment of Lumbar Disc Herniation with Radiculopathy (2012)57 Quality Rating: 5 out of 7

Discectomy is suggested to provide more effective symptom relief than medical/interventional care for patients with lumbar disc herniation with radiculopathy whose symptoms warrant surgical intervention. In patients with less severe symptoms, surgery or medical/interventional care appear to be effective for both short- and long-term relief.

3 RCTs 2 prospective comparative cohort studies

Grade: Bf

Surgical intervention prior to 6 months is suggested in patients with symptomatic lumbar disc herniation whose symptoms are severe enough to warrant surgery. Earlier surgery (within 6 months to 1 year) is associated with faster recovery and improved long-term outcomes.

4 studies (unclear study design)

Grade: Bf

The performance of surgical decompression is suggested to provide better medium-term (1 to 4 years) symptom relief as compared with medical/interventional management of patients with radiculopathy from lumbar disc herniation whose symptoms are severe enough to warrant surgery.

3 RCTs 1 prospective comparative cohort study

Grade: Bf

Surgical decompression provides long-term (greater than four years) symptom relief for patients with radiculopathy from lumbar disc herniation whose symptoms warrant surgery. It should be noted that a substantial portion (23-28%) of patients will have chronic back or leg pain.

1 retrospective comparative cohort study 5 retrospective case series

Level of Evidence: IVf

When surgery is indicated, performance of sequestrectomy or aggressive discectomy is recommended for decompression in patients with lumbar disc herniation with radiculopathy since there is no difference in rates of reherniation.

1 RCT 1 prospective comparative cohort study

Grade: Bf

Use of an operative microscope is suggested to obtain comparable outcomes to open discectomy for patients with lumbar disc herniation with radiculopathy whose symptoms warrant surgery.

2 RCTs Grade: Bf

(continued)






North American Spine Society (continued)

Endoscopic percutaneous discectomy is suggested for carefully selected patients to reduce early postoperative disability and reduce opioid use compared with open discectomy in the treatment of patients with lumbar disc herniation with radiculopathy.

3 RCTs Grade: Bf

Endoscopic percutaneous discectomy may be considered for the treatment of lumbar disc herniation with radiculopathy.

3 RCTs 4 retrospective case series

Grade: Cf

Automated percutaneous discectomy may be considered for the treatment of lumbar disc herniation with radiculopathy.

2 RCTs 4 prospective case series

Grade: Cf

In a select group of patients automated percutaneous lumbar discectomy (APLD) may achieve equivalent results to open discectomy, however, this equivalence is not felt to be generalizable to all patients with lumbar disc herniation with radiculopathy whose symptoms warrant surgery.

3 RCTs Level of Evidence: II/IIIf

(continued)






There is insufficient evidence to make a recommendation for or against the following: Urgent surgery for patients with motor deficits Use of spinal manipulation as an alternative to discectomy The specific surgical approach for far lateral disc herniation Use of tubular discectomy compared with open discectomy Use of medial facetectomy with discectomy Use of fusion for specific patient populations with lumbar disc herniation and radiculopathy Use of percutaneous electrothermal disc decompression Use of intradiscal high-pressure saline injection Use of automated percutaneous discectomy compared with open discectomy Use of plasma disc decompression/nucleoplasty Use of plasma disc decompression as compared with transforaminal epidural steroid injections in patients with lumbar disc herniation who have previously failed transforaminal epidural steroid injection therapy

-- Grade: If

American College of Occupational and Environmental Medicine Low back disorders. In occupational medicine practice guidelines: evaluation and management of common health problems and functional recovery in workers (2016)54 Quality Rating: Unknowng

Patients with evidence of specific nerve root compromise confirmed by appropriate imaging studies may be expected to potentially benefit from surgery. Quality evidence indicates that patient outcomes are not adversely affected by delaying nonemergent surgery for weeks or a few months and continued conservative care is encouraged in patients with stable or improving deficits who desire to avoid surgery. However, patients with either moderate to severe neurological deficits that are not improving or trending to improvement at 4 to 6 weeks may benefit from earlier surgical intervention. Those with progressive neurological deficit(s) are believed to have indications for immediate surgery. Those with severe deficits that do not rapidly improve are also candidates for earlier testing and referrals.

Unknowng Unknowng

(continued)






National Institute for Health and Care Excellence (United Kingdom) Percutaneous transforaminal endoscopic lumbar discectomy for sciatica: Interventional procedures guidance [IPG 556] (2016)]58 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous transforaminal endoscopic lumbar discectomy for sciatica is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit. Percutaneous transforaminal endoscopic lumbar discectomy for sciatica is a procedure that needs particular experience. Surgeons should acquire the necessary expertise through specific training and mentoring. It should only be done by surgeons who do the procedure regularly.

1 SR of observational studies 1 retrospective comparative cohort study 2 prospective case series 5 retrospective case series

None provided

National Institute for Health and Care Excellence (United Kingdom) Percutaneous interlaminar endoscopic lumbar discectomy for sciatica: Interventional procedures guidance[IPG555](2016)59 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous interlaminar endoscopic lumbar discectomy for sciatica is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit. Percutaneous interlaminar endoscopic lumbar discectomy for sciatica is a procedure that needs particular experience. Surgeons should acquire the necessary expertise through specific training and mentoring. It should only be done by surgeons who do the procedure regularly.

2 RCTs 2 retrospective comparative cohort studies 4 retrospective case series

None provided

(continued)






National Institute for Health and Care Excellence (United Kingdom) Percutaneous coblation of the intervertebral disc for low back pain and sciatica Interventional procedures guidance[IPG543](2016)60 Quality Rating: 2 out of 7

Current evidence on percutaneous coblation of the intervertebral disc for low back pain and sciatica raises no major safety concerns. The evidence on efficacy is adequate and includes large numbers of patients with appropriate follow-up periods. Therefore, this procedure may be used provided that normal arrangements are in place for clinical governance, consent and audit. As part of the consent process, patients should be informed that there is a range of treatment options available to them and also that further procedures may be needed.

1 SR 2 RCTs 1 case series

None provided

National Institute for Health and Care Excellence (United Kingdom) Percutaneous electrothermal treatment of the intervertebral disc annulus for low back pain and sciatica Interventional procedures guidance[IPG544](2016)61 Quality Rating: 2 out of 7

Current evidence on percutaneous electrothermal treatment of the intervertebral disc annulus for low back pain and sciatica raises no major safety concerns. The evidence on efficacy is inconsistent and of poor quality. Therefore, this procedure should only be used with special arrangements for clinical governance, consent and audit or research.

1 SR 1 RCT 1 Cohort study

None provided

National Institute for Health and Care Excellence (United Kingdom) Percutaneous intradiscal radiofrequency treatment of the intervertebral disc nucleus for low back pain. Interventional procedures guidance[IPG545] (2016)62 Quality Rating: 2 out of 7

Current evidence on percutaneous intradiscal radiofrequency treatment of the intervertebral disc nucleus for low back pain raises no major safety concerns. The evidence on its efficacy is limited in quantity and quality. Therefore, this procedure should only be used with special arrangements for clinical governance, consent and audit or research.

1 RCT 1 nonrandomized CT 2 case series

None provided

(continued)






National Institute for Health and Care Excellence (United Kingdom) Epiduroscopic lumbar discectomy through the sacral hiatus for sciatica

Interventional procedures guidance[IPG570] (2016)63 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of epiduroscopic lumbar discectomy through the sacral hiatus for sciatica is limited in quantity and quality. Therefore, this procedure should only be used in the context of research.

1 Cohort study None provided

National Institute for Health and Care Excellence (United Kingdom) Percutaneous intradiscal laser ablation in the lumbar spine. Interventional procedures guidance[IPG357] (2010)64 Quality Rating: 2 out of 7

Current evidence on the safety and efficacy of percutaneous intradiscal laser ablation in the lumbar spine is adequate to support the use of this procedure provided that normal arrangements are in place for clinical governance, consent and audit. Patients selected for the procedure should be limited to those with severe pain refractory to conservative treatment, in whom imaging studies show bulging of an intact disc, and who do not have neurological deficit requiring surgical decompression.

1 RCT 2 Cohort studies 2 Case series

None provided

National Institute for Health and Care Excellence (United Kingdom) Automated percutaneous mechanical lumbar discectomy: Interventional procedures guidance[IPG141])(2005)65 Quality Rating: 2 out of 7

Current evidence suggests that there are no major safety concerns associated with automated percutaneous mechanical lumbar discectomy. There is limited evidence of efficacy based on uncontrolled case series of heterogeneous groups of patients, but evidence from small randomized controlled trials shows conflicting results. In view of the uncertainties about the efficacy of the procedure, it should not be used without special arrangements for consent and for audit or research. Clinicians wishing to undertake automated percutaneous mechanical lumbar discectomy should take the following actions. Inform the clinical governance leads in their Trusts.

3 RCTs 5 case series

None provided

(continued)






National Institute for Health and Care Excellence (United Kingdom) (continued)

Ensure that patients understand the uncertainty about the procedure's efficacy and provide them with clear written information. In addition, use of the Institute's information for the public is recommended. Audit and review clinical outcomes of all patients having automated mechanical percutaneous lumbar discectomy.

National Institute for Health and Care Excellence (United Kingdom) Endoscopic laser foraminoplasty. Interventional procedures guidance[IPG31] (2003)66 Quality Rating: 2 out of 7

Current evidence of the safety and efficacy of endoscopic laser foraminoplasty does not appear adequate to support the use of this procedure without special arrangements for consent and for audit or research. Clinicians wishing to undertake endoscopic laser foraminoplasty should inform the clinical governance leads in their Trusts. They should ensure that patients offered the procedure understand the uncertainty about its safety and efficacy and should provide them with clear written information. Use of the Institute's information for the public is recommended. Clinicians should ensure that appropriate arrangements are in place for audit or research. Further research into safety and efficacy outcomes will be useful in reducing the current uncertainty. NICE is not undertaking further investigation at present.

3 Cohort studies 2 Case series

None provided

a We assessed the quality of guideline using the Appraisal of Guidelines For Research & Evaluation II (AGREE II) Instrument, version 2017.20 The lowest quality score possible is

1, the highest possible quality score is 7.

b Only recommendations from the guideline pertinent to surgical interventions for lumbar radiculopathy are summarized.

c Level B= “The panel recommends that clinicians consider offering the intervention to eligible patients”; Good= “evidence includes consistent results from well-designed, well-

conducted studies in representative populations that directly assess effects on health outcomes (at least 2 consistent higher-quality trials)

d One included trial was for treatment of sciatica with spinal stenosis, the rest were for treatment of lumbar radiculopathy

e Based on GRADE.

f Level 1=high quality RCTs or SRs of RCTs; Level II=lesser quality RCTs, prospective comparative studies, SRs that include Level II studies; Level III=Case control or

retrospective cohort studies, SRs of Level III studies, Level 4=case series; Level 5= Expert Opinion, Grade A=Good evidence (Level 1 studies with consistent findings); Grade

B=Fair evidence (Level II or III studies with consistent findings), Grade C=Poor evidence (Level IV or V studies); Grade I=insufficient or conflicting evidence not allowing a

recommendation

g The complete guideline is not publicly accessible; thus, a full quality appraisal and summary of the evidence base and strength of evidence ratings were not possible.



4. Discussion

4.1 Summary of the Evidence

Evidence maps summarizing the overall findings and strength of evidence are provided in

Figures 3, 4, 5, 6, and 7. With few exceptions, most findings were based on evidence graded as

low to very low certainty.

Surgery reduces pain and improves function more than conservative management in the short

and medium term (up to 26 weeks) but this difference does not persist in the long term (Figure

3). Surgery-related complications are rare and surgery may be cost-effective depending on a

decision-makers willingness to pay threshold.

With few exceptions, minimally-invasive surgery and standard surgery similarly reduce pain and

improve function (Figure 4). However, minimally-invasive surgery results in a quicker return to

work and lower incidence of persistent opioid use compared to standard surgery. Surgical

morbidity and reoperations are similar between both approaches. The evidence on cost-

effectiveness for minimally invasive surgery compared to standard approaches is inconclusive.

Microdiscectomy and discectomy are comparable with respect to efficacy and safety, but

microdiscectomy costs are higher (Figure 5).

For persons with recurrent lumbar radiculopathy, efficacy and safety outcomes are similar for


use is less with spinal cord stimulation (Figure 6). In contrast, surgical morbidity, persistent

opioid use and return to work outcomes are more favorable with revision endoscopic discectomy

compared to standard microdiscectomy (Figure 7).



Figure 3. Evidence map of surgery compared with nonsurgical interventions for treatment of symptomatic lumbar radiculopathy

Note: Outcomes related to surgical mortality (k=5), surgical morbidity (k=6), and reoperations (k=5) were synthesized for the surgical intervention group only as they are not appropriate for comparative evaluation with a nonsurgical intervention group. See Section 3.3.2.1, 3.3.2.2, and 3.3.2.3 for details.



Figure 4. Evidence map of minimally invasive surgery compared with discectomy or microdiscectomy for treatment of symptomatic lumbar radiculopathy



Figure 5. Evidence map of microdiscectomy compared with discectomy for treatment of symptomatic lumbar radiculopathy



Figure 6. Evidence map of repeat lumbosacral decompression compared with spinal cord stimulation for treatment of recurrent symptomatic lumbar radiculopathy



Figure 7. Evidence map of revision endoscopic discectomy compared with revision microdiscectomy for treatment of recurrent symptomatic lumbar radiculopathy



4.2 Limitations of the Evidence Base

The evidence we identified for inclusion in this HTA has several limitations.

4.2.1 High risk of bias among included studies

We rated nearly half of included studies as having a high risk of bias. Some sources of bias

across included studies were common; for example, all but one study did not blind participants,

caregivers, or clinicians to treatment allocation and most did not blind outcome assessors.

Knowledge of treatment allocation has the potential to influence other decisions about care that

may be related to the outcome, and also to influence the outcome assessment itself. Because

most all studies use patient-reported outcomes, this introduces some concerns for bias. Although

blinding treatment allocation can be very challenging to perform in trials of surgery, particularly

those comparing surgical interventions to nonsurgical interventions, the risk of bias nonetheless

remains and should be acknowledged. The direction of bias from nonblinding largely depends on

the beliefs and attitudes of participants, clinicians, and outcome assessors, so cannot always be

predicted.

In contrast, deviations from the intended intervention because of crossovers and contamination

were the source of bias contributing to high risk of bias ratings for some studies. In an intent-to

treat analysis, which preserves randomization and thus mitigates other sources of bias, the

direction of bias from crossovers is predictable and results in bias toward a null effect. Thus,

estimates from the intent-to-treat analysis are conservative, and may underestimate the effect

when deviations from the intended intervention occur. This is evident with respect to pain and

functional outcomes for surgery compared to non-surgical interventions; the evidence favored

surgery in the short- and medium-term, but these differences in the long-term may have been

mitigated by cumulative crossovers that occurred over time. Because of extensive crossovers,

several included studies also reported ‘as treated’ analyses. Participants are included in these

analyses according to the treatment they actually received (as opposed to the treatment to which

they were randomly allocated). Weinstein et al.[SPORT] reported an as-treated analysis in

addition to the intent-to-treat analysis and found much larger favorable effects for

discectomy/microdiscectomy compared with conservative management through 2 years of

followup.21 The between-group difference at 52 weeks for the SF-36 bodily pain subscale was

15.0 ((5 % CI, 10.9 to 19.2), the SF-36 physical functioning subscale 17.5 (95% CI, 13.6 to

21.5), and the Oswestry disability index (-15.0 (95% CI, -18.3 to -11.7). Although as-treated

analyses may offer some insight into the magnitude of bias toward the null effect, because

participants generally do not cross over at random but for reasons that are also related to the

outcomes (e.g., pain, symptoms), these analyses can introduce other biases on the effect estimate.

Lastly, the sources of bias contributing to high risk of bias ratings for a few studies was

inadequate randomization (e.g., using even/odd sequences) or inadequate allocation

concealment, high attrition (e.g., loss to followup and missing data).

4.2.3 Studies generally underpowered for many outcomes of interest

Only 11 of the 24 included RCTs for efficacy and safety designated a primary outcome and

described the sample size required to detect an a priori effect size. Few described how this effect

size was determined or whether it represented a minimally important clinical difference. Eight

studies were powered based on pain or function outcomes, three studies were powered based on



duration of surgery or hospital stay, and one study was powered based on differences in success.

Consequently, study samples were not adequate to detect many of the efficacy outcomes that

were reported, or safety outcomes, which occurred at a low frequency. Thus, effect estimates

were often imprecise, which resulted in downgrading the strength of evidence ratings from

moderate to low, or from low to very low.

4.2.4 Variation in diagnosis and severity of symptoms

Most studies required participants to have a clinical diagnosis of radiculopathy confirmed by

imaging (usually CT or MRI) for enrollment. However, few studies described the criteria for

clinical diagnosis. Further, the duration of symptoms and criteria related to provision of

conservative therapy prior to enrollment was variable across studies. The duration between

enrollment and receiving surgery was also variable, in some cases months.

4.2.5 Limited number of comparative effectiveness trials for any given

procedure

We identified 15 trials comparing minimally invasive surgery to open surgery. However, most of

these interventions were only evaluated by 1 to 3 RCTs and variations in the outcomes reported

limited our ability to draw conclusions for specific minimally invasive procedures. Further,

variation in nomenclature for these procedures and lack of detail regarding the procedures may

have also limited out ability to synthesize findings for specific interventions. In addition, many

studies lacked a full description of the surgical intervention, including the procedure, the skill

experience of the surgeon and surgical team, and pre- and postoperative care.

4.2.6 Variation in type, timing, and completeness in reporting outcomes

Some studies reported between-group differences at multiple follow-up time points without a

priori specification of a primary time point; others more appropriately used repeated measures

analysis, to account for multiple observations over time, and some reported both. Our ability to

conduct quantitative syntheses (i.e., meta-analysis) was limited by variation in specific outcomes

reported and by incomplete reporting. For efficacy outcomes, studies used a variety of pain and

function measures, and measured outcomes at different follow-up time points. Further, some

studies only reported adjusted difference-in-difference treatment effects, while other studies

reported values only at followup, not adjusted for baseline values. Some studies only reported

short-term outcomes, others only reported long-term outcomes. Some studies reported results of

statistical significance testing for between-group differences, but did not report actual outcome

values. Others reported actual outcome values, but no measures of variance. Combined, all of

these data reporting issues limited the extent to which we could conduct quantitative synthesis.

4.2.7 Limited number of United States cost studies

Whereas efficacy and safety outcomes from studies conducted outside of the United States are

likely applicable to U.S. settings, it is not clear cost studies conducted based on RCTs outside of

the United States would apply to U.S. settings. The only RCT conducted in the United States

included in this HTA reported direct medical costs in the surgical group that were nearly triple

the health care costs in any of the other cost studies that reported health care costs separately.50

Although the effectiveness inputs from non-U.S. studies used in cost-effectiveness analyses are

likely applicable, the extreme differences in how health care services are organized and financed



between U.S. and non-U.S. countries probably reduces the applicability of the cost inputs used in

non-U.S. studies.

4.3 Other related HTAs

The only related HTA that we identified was commissioned by the National Institute for Health

Research (U.K.) Health Technology Assessment programme.67 This HTA included both surgical

and nonsurgical interventions for the management of sciatica and used a network meta-analysis

to provide a measure of relative therapeutic effect across 18 different treatment categories. The

findings suggest that nonopioid medication, epidural corticosteroids injections, and disc surgery

are effective for reducing sciatica This HTA also concluded that stepped care approaches to

treatment are cost-effective relative to direct referral for surgery.

4.4 Selected payer coverage policies

The Centers for Medicare and Medicaid Services (CMS) does not have a national coverage

determination related to open standard or microsurgical decompressive procedures (i.e.,

discectomy, microdiscectomy, foraminotomy, laminectomy/otomy). CMS recognizes the use of

lasers to alter, revise, or destroy tissues in place of more conventional techniques as part of a

surgical procedure. Medicare administrative contractors have been advised to use discretion in

determining coverage for procedures performed with a laser when the laser has been FDA-

approved, the procedure is considered reasonable and necessary, and a noncoverage instruction

does not exist (effective date May 1, 1997).68 CMS does have a national coverage determination

related to thermal intradiscal procedures; these procedures are not covered (effective date

January 1, 2009).68 Percutaneous disc decompression falls within the category of procedures

covered by this determination. Table 44 provides an overview of other payer coverage policies

and Table 45 summarizes excerpts from these policies that are relevant to surgery for lumbar

radiculopathy and disc herniation.

Table 44. Overview of payer coverage policies

Procedure Medicare Premera Regence Cigna United Aetna Humana Kaiser

Laminectomy, laminotomy, discectomy, foraminotomy (open technique including microsurgical approaches)

-- a -- -- -- a a --

Automated percutaneous lumbar disc decompression

b c --

(Percutaneous) endoscopic discectomy


--

(Percutaneous) laser discectomy


--

Percutaneous nucleoplasty with coblation technology

-- --

= covered; = not covered; — = no policy identified

a If specific clinical criteria are met. See Table 45 in Full Report for details.

b All percutaneous disc decompression procedures fall under a Medicare National Coverage Determination related to thermal

intradiscal procedures.

c Also covers percutaneous manual discectomy, see Table 45 for details.



Table 45. Selected payer coverage for surgery for lumbar radiculopathy

Payer; Effective Date Policy

Premera (Blue Cross)108-110

July 1, 2017 July 1, 2017 July 1, 2017 July 1, 2017 April 1, 2017

Premera may consider lumbar spine decompression surgery (discectomy, foraminotomy, laminotomy) medically necessary for the rapid (48 hours or less) progression of neurologic impairment (e.g., cauda equina syndrome, foot drop, extremity weakness, saddle anesthesia, sudden onset of bladder or bowel dysfunction); or in the absence of rapid progression when all of the following criteria are met:

All other sources of low back pain have been ruled out AND

MRI or CT with myelogram within the past 12 months shows nerve root compression that corresponds to symptoms and physical examination findings or there is definitive neurological localization by other means AND

Persistent, debilitating pain radiating from the low back down to the lower extremity is present daily and limits activities of daily living AND

Neurological deficits (e.g., reflex change in the legs, dermatomal sensory loss, motor weakness) or alternative signs of lumbar root irritation (e.g., positive leg raising test) are present on physical examination AND

The patient has failed at least 6 weeks of conservative therapy such as activity modification, oral analgesics/anti-inflammatories, physical therapy, chiropractic manipulation, epidural steroid injections.

Lumbar laminectomy may also be considered medically necessary for the rapid progression of neurologic impairment or when criteria related to the presence of lumbar spinal stenosis are met. Lumbar spine decompression surgery is considered not medically necessary when no clinical indication is documented and there are no confirmatory physical and radiologic findings that meet the relevant criteria listed above. The provider’s choice of interventional surgery depends on the specific member’s symptoms and imaging findings. Automated percutaneous and percutaneous endoscopic discectomy are considered investigational as techniques for intervertebral disc decompression in patients with back pain and/or radiculopathy related to disc herniation in the lumbar, thoracic, or cervical spine. Decompression of the intervertebral disc using laser energy (laser discectomy) or radiofrequency coblation (nucleoplasty) are considered investigational as techniques of disc decompression and treatment of associated pain.

Regence (Blue Shield)111,112 August 1, 2017

Automated percutaneous and percutaneous endoscopic discectomy are considered investigational as techniques for intervertebral disc decompression in patients with back pain and/or radiculopathy related to disc herniation in the lumbar, thoracic, or cervical spine. Decompression of intervertebral discs using laser energy (laser discectomy) or radiofrequency energy (nucleoplasty) are considered investigational for all indications, including but not limited to disc decompression and treatment of associated pain.

(continued)



Table 45. Selected payer coverage for surgery for lumbar radiculopathy (continued)


Aetna January 17, 2018

Aetna considers lumbar decompression with or without discectomy medically necessary for rapid progression of neurological impairment (e.g., foot drop, extremity weakness, numbness or decreased sensation, saddle anesthesia, bladder dysfunction or bowel dysfunction) confirmed by imaging studies (e.g., CT or MRI). Aetna considers lumbar laminectomy medically necessary for individuals with a herniated disc when all of the following criteria are met: All other reasonable sources of pain have been ruled out; and

Central/lateral recess or foraminal stenosis graded as moderate, moderate to severe or severe (not mild or mild to moderate); and

Imaging studies (e.g., CT or MRI) indicate nerve root compression, that corresponds to the clinical findings of the specific affected nerve root; and

Member has failed at least 6 weeks of conservative therapy; and Member's activities of daily living are limited by persistent pain radiating from the back down to the lower extremity; and Presence of neurological abnormalities (e.g., reflex change, positive straight leg raising, sensory loss, weakness) persist on examination and correspond to the specific affected nerve root. Aetna considers percutaneous lumbar discectomy, manual or automated, medically necessary for treatment of herniated lumbar discs when all of the following are met: Member is otherwise a candidate for open laminectomy; and Member has failed 6 months of conservative treatment; and Diagnostic studies show that the nuclear bulge of the disc is contained within the annulus (i.e., the herniated disc is contained); and Member has no previous surgery or chemonucleolysis of the disc to be treated; and Member must have typical clinical symptoms of radicular pain corresponding to the level of disc involvement. Aetna considers the following procedures experimental and investigational: Endoscopic disc decompression, ablation, or annular modulation using the DiscFX System; Endoscopic laser foraminoplasty, endoscopic foraminotomy, laminotomy, and rhizotomy (endoscopic radiofrequency ablation) Endoscopic transforaminal discectomy Far lateral microendoscopic discectomy (FLMED) for extra-foraminal lumbar disc herniations or other indications; Far lateral microendoscopic discectomy (FLMED) for extra-foraminal lumbar disc herniations or other indications; Far lateral microendoscopic discectomy (FLMED) for extra-foraminal lumbar disc herniations or other indications; Further Reimbursement Notes: Laser: Clinical studies have not established a clinically significant benefit of use of a laser over a scalpel in spinal surgery. No additional benefit will be provided for the use of a laser in spinal surgery. Microscope and endoscope: Use of a microscope or endoscope is considered an integral part of the spinal surgery and not separately reimbursable.

(continued)





United113 August 1, 2017

United considers percutaneous discectomy and decompression procedures as unproven and not medically necessary for treating discogenic pain, including, but are not limited to, the following procedures: Nucleoplasty [percutaneous disc decompression or percutaneous plasma discectomy] Laser discectomy [laser disc decompression; laser-assisted disc decompression (LADD); or percutaneous endoscopic discectomy, with or without laser] Yeung endoscopic spinal surgery [arthroscopic microdiscectomy or percutaneous endoscopic discectomy] Transforaminal and/or interlaminar [transforaminal and interlaminar approach]

Cigna114 June 15, 2017

Cigna considers percutaneous, endoscopic laminectomy and disc decompression procedures of the lumbar spine experimental, investigational, and unproven when used to report: Automated percutaneous lumbar discectomy/automated percutaneous nucleotomy Endoscopic anterior spinal surgery/Yeung endoscopic spinal system/percutaneous endoscopic discectomy/arthroscopic microdiscectomy, selective endoscopic discectomy Endoscopic disc decompression, ablation, or annular modulation using the DiscFX™ System Percutaneous laminotomy/laminectomy, percutaneous spinal decompression (e.g., mild® Procedure) Percutaneous laser discectomy /decompression, laser-assisted disc decompression Thermal intradiscal procedures are also considered experimental, investigational or unproven when used to report: Intervertebral disc biacuplasty Intradiscal electrothermal annuloplasty Percutaneous intradiscal radiofrequency thermocoagulation, intradiscal radiofrequency thermomodulation, percutaneous radiofrequency thermomodulation Coblation® Nucleoplasty™, disc nucleoplasty, decompression nucleoplasty plasma disc decompression Intraosseous radiofrequency nerve ablation of basivertebral nerve

(continued)





Humana115 September 28, 2017

Humana members may be eligible for discectomy (including microdiscectomy) for the following indications: Evidence of myelopathy, confirmed by CT or MRI, with both corresponding clinical symptoms and corresponding objective neurological signs Herniated disc, confirmed by imaging studies, when accompanied by radicular pain that has persisted despite 12 consecutive weeks of appropriate conservative treatment (e.g., rest, medications, physical therapy) Rapidly progressive neurological signs/symptoms of lumbar spine compression confirmed by imaging studies. Spinal fractures, infections, and tumors. Humana members may be eligible for a lumbar laminectomy, laminotomy, foraminectomy, foraminotomy or foraminolaminectomy for the following indications: Cauda equina syndrome (bowel or bladder dysfunction, bilateral lower extremity weakness/numbness/decreased sensation, saddle anesthesia) confirmed by imaging studies Herniated disc, foraminal stenosis or spinal stenosis at the level corresponding with clinical findings confirmed by imaging studies when accompanied by both radicular pain that has persisted despite 12 consecutive weeks of appropriate conservative treatment and physical and/or neurological abnormalities suggestive of nerve root or spinal cord compression Rapidly progressive neurologic signs/symptoms of lumbar spine compression confirmed by imaging studies Spinal fractures, infection, injury, tumor. Spondylolisthesis A minimally invasive approach (e.g., endoscopic) which allows direct visualization of the surgical field and anatomy, is integral to the procedure and is not separately reimbursable. This applies to microendoscopic discectomy, tubular microdiscectomy, and other systems designed for minimally invasive procedures. Humana members may NOT be eligible for other types of discectomy procedure including, but not limited to, the following procedures considered experimental and investigational: Accurascope DND Automated percutaneous lumbar discectomy, including but not limited to the Stryker Dekompressor lumbar discectomy probe. Laser discectomy, regardless of the approach, including percutaneous laser discectomy, laser-assisted discectomy, laser disc decompression, laser-assisted disc decompression or percutaneous laser disc decompression Percutaneous discectomy techniques not previously listed including, but not limited to, the HydroCision/HydroDiscectomy Humana members may NOT be eligible for other types of laminectomy, laminotomy, foraminectomy, foraminotomy, foraminolaminectomy, laminoplasty, corpectomy or decompression procedure including, but not limited to, the following procedures considered experimental and investigational: ANY percutaneous laminectomy, laminotomy, foraminectomy, foraminotomy, foraminolaminectomy, laminoplasty or corpectomy Laser laminectomy Percutaneous image-guided lumbar decompression including, but not be limited to, the MILD procedure and the Totalis Direct Decompression system

(continued)





Kaiser Permanente

No policies specific to the coverage of surgeries for lumbar radiculopathy.

Medicaid116 Policies vary by state; example state Medicaid policy for North Carolina: Medicaid and North Carolina Health Check shall cover lumbar decompression surgery (discectomy, microdiscectomy, corpectomy, hemicorpectomy, foraminectomy, foraminoplasty, foraminotomy, laminectomy, hemilaminectomy, laminotomy, laminoplasty, and osteophytectomy) when all other reasonable sources of pain have been ruled out and the beneficiary meets the one or more of following specific criteria: Rapidly progressive neurological findings of nerve root or spinal cord compression, with imaging evidence of pathology that correlates with clinical findings (with or without gait or sphincter disturbance); Elective surgery needed as indicated by all the following when the beneficiary has failed at least six (6) consecutive weeks of conservative medical management (unless imaging indicates the need for urgent intervention): A Herniated disc with all of the following: i. Nerve or spinal cord impingement seen on imaging studies; ii. Clinical findings consistent with impingement; and iii. All major psychosocial and substance use issues have been addressed.

b Persistent pain and symptoms or findings that have not improved after at least six (6) consecutive weeks of conservative medical management, consisting of one or more of the following: i. Severe disabling radiculopathy; or ii. Clinical findings of nerve root compromise; 3) Spinal stenosis, spondylolisthesis, spinal fracture, or cauda equina syndrome.

Abbreviations: CT = computed tomography; DND = discectomy and neural decompression procedure; MRI = magnetic resonance imaging.



In general, payers cover decompressive procedures, including discectomy, laminectomy/otomy,

foraminectomy/otomy, including microsurgical approaches, for disc herniation with radicular

symptoms. Specific criteria vary by payer but often include a failed trial of conservative

management for 6 to 12 weeks. Most payers also require imaging confirmation of nerve root

compression that corresponds to symptoms and physical examination findings. The coverage of

minimally invasive procedures varies by payer.

4.5 Limitations of this HTA

This HTA was limited to studies and other information published or publicly available in

English. The electronic search was limited to only three databases. For efficiency, we relied on

hand searches of existing systematic reviews to identify eligible studies published prior to 2007.

Although this approach may have resulted in missed studies, we think this is unlikely since we

hand searched more than 40 systematic reviews. We used a single reviewer to screen titles and

abstract; however, we mitigated this risk through reviewer training, quantitative assessment of

interrater reliability during initial dual-review of 50 titles/abstracts, and using a low threshold for

reviewers to request a second screening by another team member. We only included efficacy

outcomes reported at 4 weeks or later; thus, immediate and very-short term benefits are not

reflected in our synthesis.

4.6 Ongoing Research and Future Research Needs

We did not identify any ongoing trials of surgical interventions specifically for lumbar

radiculopathy through our search of the U.S. clinical trials registry (clinicaltrials.gov). Several

trials are ongoing related injections of biologics (e.g., condoliase into nucleus pulposus)69 or

pharmacologics (e.g., epidural clonidine)70 or use of adjunctive treatments (e.g., epidural steroid

injections, stem cell injections, annular repair technologies) during or after discectomy to

improve outcomes. The challenges faced in conducting methodologically rigorous randomized

trials of surgical interventions are well-documented.71 However, additional trials on treatment of

lumbar radiculopathy with the same methodologic flaws will be unlikely to change the certainty

of findings. Additional research on patient preferences and values related to timing of treatment

or surgery, and establishment of minimally important clinical differences in outcomes that are

specific to sciatica would also advance research in this area Finally, advanced analytic and

statistical techniques could be used within trials to quantify and mitigate the impact of crossovers

on treatment effects and could be used within observational studies to mitigate biases introduced

by nonrandomized study designs, potentially broadening the evidence base available to address

important research questions.

5. Conclusion

Surgery for symptomatic lumbar radiculopathy reduced pain and improved physical function

more in the short and medium term (up to 26 weeks) compared to non-surgical interventions, but

these findings did not persist at one year or longer followup. Surgery compared with non-

surgical interventions resulted in similar improvements in neurologic symptoms, quality of life,

and return to work. No surgical deaths were reported by included studies and no differences in

persistent opioid use or all-cause mortality were observed. Surgery may be cost-effective relative

to non-surgical interventions depending on a decision maker’s willingness to pay threshold. All

findings were based on evidence graded as low to very low certainty.



Minimally invasive surgery was comparable to microdiscectomy or discectomy for reducing pain

and improving function in the short and long term based on most outcome measures reported.

Surgical morbidity and reoperations were similar in most studies, but a few studies suggested a

lower frequency of reoperation among participants that received standard surgery. Persistent

opioid use was lower among participants that received minimally-invasive surgery. The evidence

on cost-effectiveness of minimally invasive surgery is inconclusive. All findings were based


Microdiscectomy and discectomy were comparable with respect to efficacy and safety, but

microdiscectomy costs were higher. Efficacy and safety findings were based on evidence graded

as low to very low certainty; cost-effectiveness was based on evidence graded as moderate

certainty.

For persons with recurrent lumbar radiculopathy, efficacy and safety outcomes were similar for


use was less with spinal cord stimulation. In contrast, persistent opioid use and return to work

outcomes were more favorable with revision endoscopic discectomy compared to standard

microdiscectomy. All but one finding (surgical mortality) was based on evidence graded as very

low certainty.



6. References

1. Konstantinou K, Dunn KM. Sciatica: review of epidemiological studies and prevalence

estimates. Spine (Phila Pa 1976). 2008;33(22):2464-2472. PMID: 18923325. doi:

10.1097/BRS.0b013e318183a4a2

2. Rhee JM, Schaufele M, Abdu WA. Radiculopathy and the herniated lumbar disk:

controversies regarding pathophysiology and management. Instr Course Lect.

2007;56:287-299. PMID: 17472314.

3. Tarulli AW, Raynor EM. Lumbosacral radiculopathy. Neurol Clin. 2007;25(2):387-405.

PMID: 17445735. doi: 10.1016/j.ncl.2007.01.008

4. Casey E. Natural history of radiculopathy. Phys Med Rehabil Clin N Am. 2011;22(1):1-5.

PMID: 21292142. doi: 10.1016/j.pmr.2010.10.001

5. Weber H. Lumbar disc herniation. A controlled, prospective study with ten years of

observation. Spine (Phila Pa 1976). 1983;8(2):131-140. PMID: 6857385.

6. Heliövaara M, MÄkelÄ M, Knekt P, Impivaara O, Aromaa A. Determinants of sciatica

and low-back pain. Spine. 1991;16(6):608-614. PMID: 1830689.

7. Varlotta G, Brown M, Kelsey J, Golden A. Familial predisposition for herniation of a

lumbar disc in patients who are less than twenty-one years old. JBJS. 1991;73(1):124-

128. PMID: 1824705.

8. St. Francis Medical Technologies, Inc. Draft Package Insert: S STOP Interspinous

Process Decompression (IPD) System. Published 2005. Updated November 9. Accessed

November 8, 2017.

9. Paradigm Spine. Instructions for Use: coflex@ Interlaminar Technology.

https://www.accessdata.fda.gov/cdrh_docs/pdf11/P110008C.pdf. Published n.d. Accessed

November 8, 2017.

10. X-Sten Corp. Premarket Notification (51 0(k)] Summary.

https://www.accessdata.fda.gov/cdrh_docs/pdf6/K062038.pdf Published 2006. Accessed

November 8, 2017.

11. Vavken P, Ganal-Antonio AK, Quidde J, Shen FH, Chapman JR, Samartzis D.

Fundamentals of Clinical Outcomes Assessment for Spinal Disorders: Clinical Outcome

Instruments and Applications. Global Spine J. 2015;5(4):329-338. PMID: 26225283. doi:

10.1055/s-0034-1396046

12. Department of Health and Human Services. Patient-Reported Outcomes Measurement

Information System (PROMIS). http://www.healthmeasures.net/explore-measurement-

systems/promis. Published 2017. Accessed November 8, 2017.

13. United Nations Development Programme. Human Development Index. Website.

http://hdr.undp.org/en/composite/HDI. Published 2016. Accessed October 15, 2017.

14. Higgins JPT, Sterne JAC, Savović J, et al. A revised tool for assessing risk of bias in

randomized trials. In: Chandler J, McKenzie J, Boutron I, Welch V, eds. Cochrane

Methods. Vol 102016. https://sites.google.com/site/riskofbiastool/welcome/rob-2-0-tool

15. Bureau of the Fiscal Service. Treasury Reporting Rates of Exchange.

https://www.fiscal.treasury.gov/fsreports/rpt/treasRptRateExch/historicalRates.htm.

Published 2017. Updated October 12. Accessed October 19, 2017.

16. United States Department of Labor, Bureau of Labor Statistics. Consumer Price Index.

https://www.bls.gov/cpi/. Published n.d. Accessed October 19, 2017.

https://www.accessdata.fda.gov/cdrh_docs/pdf11/P110008C.pdf

https://www.accessdata.fda.gov/cdrh_docs/pdf6/K062038.pdf

http://www.healthmeasures.net/explore-measurement-systems/promis

http://www.healthmeasures.net/explore-measurement-systems/promis

http://hdr.undp.org/en/composite/HDI

https://sites.google.com/site/riskofbiastool/welcome/rob-2-0-tool

https://www.fiscal.treasury.gov/fsreports/rpt/treasRptRateExch/historicalRates.htm

https://www.bls.gov/cpi/



17. Viechtbauer W. Conducting Meta-Analyses in R with the metafor Package. Journal of

Statistical Software. 2010;36(3):1-48. PMID: WOS:000281593200001.

18. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating

quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926.

PMID: 18436948. doi: 10.1136/bmj.39489.470347.AD

19. Brouwers MC, Kho ME, Browman GP, et al. AGREE II: Advancing guideline

development, reporting, and evaluation in healthcare. CMAJ. 2017(182):E839-842.

20. AGREE Next Steps Consortium. The AGREE II Instrument [Electronic Version]. 2017.

http://www.agreetrust.org. Accessed January 15, 2018.

21. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical vs nonoperative treatment for

lumbar disk herniation: the Spine Patient Outcomes Research Trial (SPORT): a

randomized trial. Jama. 2006;296(20):2441-2450. PMID: 17119140. doi:

10.1001/jama.296.20.2441

22. McMorland G, Suter E, Casha S, du Plessis SJ, Hurlbert RJ. Manipulation or

microdiskectomy for sciatica? A prospective randomized clinical study. J Manipulative

Physiol Ther. 2010;33(8):576-584. PMID: 21036279. doi: 10.1016/j.jmpt.2010.08.013

23. Huang TJ, Hsu RW, Li YY, Cheng CC. Less systemic cytokine response in patients

following microendoscopic versus open lumbar discectomy. J Orthop Res.

2005;23(2):406-411. PMID: 15734255. doi: 10.1016/j.orthres.2004.08.010

24. Sasaoka R, Nakamura H, Konishi S, et al. Objective assessment of reduced invasiveness

in MED. Compared with conventional one-level laminotomy. Eur Spine J.

2006;15(5):577-582. PMID: 15926058. doi: 10.1007/s00586-005-0912-8

25. Tullberg T, Isacson J, Weidenhielm L. Does microscopic removal of lumbar disc

herniation lead to better results than the standard procedure? Results of a one-year

randomized study. Spine (Phila Pa 1976). 1993;18(1):24-27. PMID: 8434321.

26. Thome C, Barth M, Scharf J, Schmiedek P. Outcome after lumbar sequestrectomy

compared with microdiscectomy: a prospective randomized study. J Neurosurg Spine.

2005;2(3):271-278. PMID: 15796351. doi: 10.3171/spi.2005.2.3.0271

27. Teli M, Lovi A, Brayda-Bruno M, et al. Higher risk of dural tears and recurrent

herniation with lumbar micro-endoscopic discectomy. Eur Spine J. 2010;19(3):443-450.

PMID: 20127495. doi: 10.1007/s00586-010-1290-4

28. Ryang YM, Oertel MF, Mayfrank L, Gilsbach JM, Rohde V. Standard open

microdiscectomy versus minimal access trocar microdiscectomy: results of a prospective

randomized study. Neurosurgery. 2008;62(1):174-181; discussion 181-172. PMID:

18300905. doi: 10.1227/01.neu.0000311075.56486.c5

29. Ruetten S, Komp M, Merk H, Godolias G. Full-endoscopic interlaminar and

transforaminal lumbar discectomy versus conventional microsurgical technique: a

prospective, randomized, controlled study. Spine (Phila Pa 1976). 2008;33(9):931-939.

PMID: 18427312. doi: 10.1097/BRS.0b013e31816c8af7

30. Peul WC, van Houwelingen HC, van den Hout WB, et al. Surgery versus prolonged

conservative treatment for sciatica. N Engl J Med. 2007;356(22):2245-2256. PMID:

17538084. doi: 10.1056/NEJMoa064039

31. Osterman H SS, Malmivaara A, Karppinen J. Surgery for disc herniation. A randomized

controlled trial with 2-year follow-up. Paper presented at: Proceedings of the

International Society for Study of the Lumbar Spine. 2003.

http://www.agreetrust.org/

https://www.ncbi.nlm.nih.gov/pubmed/17119140












32. Mayer HM, Brock M. Percutaneous endoscopic discectomy: surgical technique and

preliminary results compared to microsurgical discectomy. J Neurosurg. 1993;78(2):216-

225. PMID: 8267686. doi: 10.3171/jns.1993.78.2.0216

33. Henriksen L, Schmidt K, Eskesen V, Jantzen E. A controlled study of microsurgical

versus standard lumbar discectomy. Br J Neurosurg. 1996;10(3):289-293. PMID:

8799541.

34. Franke J, Greiner-Perth R, Boehm H, et al. Comparison of a minimally invasive

procedure versus standard microscopic discotomy: a prospective randomised controlled

clinical trial. Eur Spine J. 2009;18(7):992-1000. PMID: 19360440. doi: 10.1007/s00586-

009-0964-2

35. Erginousakis D, Filippiadis DK, Malagari A, et al. Comparative prospective randomized

study comparing conservative treatment and percutaneous disk decompression for

treatment of intervertebral disk herniation. Radiology. 2011;260(2):487-493. PMID:

21613439. doi: 10.1148/radiol.11101094

36. Chatterjee S, Foy PM, Findlay GF. Report of a controlled clinical trial comparing

automated percutaneous lumbar discectomy and microdiscectomy in the treatment of

contained lumbar disc herniation. Spine (Phila Pa 1976). 1995;20(6):734-738. PMID:

7604351.

37. Brouwer PA, Brand R, van den Akker-van Marle ME, et al. Percutaneous laser disc

decompression versus conventional microdiscectomy in sciatica: a randomized controlled

trial. Spine J. 2015;15(5):857-865. PMID: 25614151. doi: 10.1016/j.spinee.2015.01.020

38. Arts MP, Brand R, van den Akker ME, Koes BW, Bartels RH, Peul WC. Tubular

diskectomy vs conventional microdiskectomy for sciatica: a randomized controlled trial.

JAMA. 2009;302(2):149-158. PMID: 19584344. doi: 10.1001/jama.2009.972

39. Gerszten PC, Smuck M, Rathmell JP, et al. Plasma disc decompression compared with

fluoroscopy-guided transforaminal epidural steroid injections for symptomatic contained

lumbar disc herniation: a prospective, randomized, controlled trial. J Neurosurg Spine.

2010;12(4):357-371. PMID: 20201654. doi: 10.3171/2009.10.spine09208

40. Haines SJ, Jordan N, Boen JR, Nyman JA, Oldridge NB, Lindgren BR. Discectomy

strategies for lumbar disc herniation: results of the LAPDOG trial. J Clin Neurosci.

2002;9(4):411-417. PMID: 12217670.

41. Hermantin FU, Peters T, Quartararo L, Kambin P. A prospective, randomized study

comparing the results of open discectomy with those of video-assisted arthroscopic

microdiscectomy. J Bone Joint Surg Am. 1999;81(7):958-965. PMID: 10428127.

42. Malter AD, Larson EB, Urban N, Deyo RA. Cost-effectiveness of lumbar discectomy for

the treatment of herniated intervertebral disc. Spine (Phila Pa 1976). 1996;21(9):1048-

1054; discussion 1055. PMID: 8724089.

43. Arts MP, Brand R, van den Akker ME, et al. Tubular diskectomy vs conventional

microdiskectomy for the treatment of lumbar disk herniation: 2-year results of a double-

blind randomized controlled trial. Neurosurgery. 2011;69(1):135-144; discussion 144.

PMID: 21792119. doi: 10.1227/NEU.0b013e318214a98c

44. Lurie JD, Tosteson TD, Tosteson AN, et al. Surgical versus nonoperative treatment for

lumbar disc herniation: eight-year results for the spine patient outcomes research trial.

Spine (Phila Pa 1976). 2014;39(1):3-16. PMID: 24153171. doi:

10.1097/BRS.0000000000000088
















45. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical versus nonoperative treatment for

lumbar disc herniation: four-year results for the Spine Patient Outcomes Research Trial

(SPORT). Spine (Phila Pa 1976). 2008;33(25):2789-2800. PMID: 19018250. doi:

10.1097/BRS.0b013e31818ed8f4

46. National Guideline Centre. National Institute for Health and Care Excellence: Clinical

Guidelines. Low Back Pain and Sciatica in Over 16s: Assessment and Management.

London: National Institute for Health and Care Excellence; 2016.

47. North RB, Kidd DH, Farrokhi F, Piantadosi SA. Spinal cord stimulation versus repeated

lumbosacral spine surgery for chronic pain: a randomized, controlled trial. Neurosurgery.

2005;56(1):98-106; discussion 106-107. PMID: 15617591.

48. Ruetten S, Komp M, Merk H, Godolias G. Recurrent lumbar disc herniation after

conventional discectomy: a prospective, randomized study comparing full-endoscopic

interlaminar and transforaminal versus microsurgical revision. J Spinal Disord Tech.

2009;22(2):122-129. PMID: 19342934. doi: 10.1097/BSD.0b013e318175ddb4

49. van den Hout WB, Peul WC, Koes BW, Brand R, Kievit J, Thomeer RT. Prolonged

conservative care versus early surgery in patients with sciatica from lumbar disc

herniation: cost utility analysis alongside a randomised controlled trial. Bmj.

2008;336(7657):1351-1354. PMID: 18502912. doi: 10.1136/bmj.39583.709074.BE

50. Tosteson AN, Skinner JS, Tosteson TD, et al. The cost effectiveness of surgical versus

non-operative treatment for lumbar disc herniation over two years: evidence from the

Spine Patient Outcomes Research Trial (SPORT). Spine. 2008;33(19):2108.

51. van den Akker ME, Arts MP, van den Hout WB, Brand R, Koes BW, Peul WC. Tubular

diskectomy vs conventional microdiskectomy for the treatment of lumbar disk-related

sciatica: cost utility analysis alongside a double-blind randomized controlled trial.

Neurosurgery. 2011;69(4):829-835; discussion 835-826. PMID: 21623248. doi:

10.1227/NEU.0b013e31822578f6

52. van den Akker-van Marle ME, Brouwer PA, Brand R, et al. Percutaneous laser disc

decompression versus microdiscectomy for sciatica: Cost utility analysis alongside a

randomized controlled trial. Interventional Neuroradiology. 2017;23(5):538-545.

53. Stevenson RC, McCabe CJ, Findlay AM. An economic evaluation of a clinical trial to

compare automated percutaneous lumbar discectomy with microdiscectomy in the

treatment of contained lumbar disc herniation. Spine (Phila Pa 1976). 1995;20(6):739-

742. PMID: 7604352.

54. National Guideline Clearinghouse. Low back disorders: American College of

Occupational and Environmental Medicine. Rockville MD: Agency for Healthcare

Research and Quality (AHRQ); 2016.

https://www.guideline.gov/summaries/summary/50687/low-back-disorders?q=lumbar

55. Chou R, Loeser JD, Owens DK, et al. Interventional therapies, surgery, and

interdisciplinary rehabilitation for low back pain: an evidence-based clinical practice

guideline from the American Pain Society. Spine (Phila Pa 1976). 2009;34(10):1066-

1077. PMID: 19363457. doi: 10.1097/BRS.0b013e3181a1390d

56. Manchikanti L, Abdi S, Atluri S, et al. An update of comprehensive evidence-based

guidelines for interventional techniques in chronic spinal pain. Part II: guidance and

recommendations. Pain Physician. 2013;16(2 Suppl):S49-283. PMID: 23615883.

57. North American Spine Society. Clinical Guidelines for Diagnosis and Treatment of

Lumbar Disc Herniation with Radiculopathy. Burr Ridge, IL: Society NAS; 2012.







https://www.guideline.gov/summaries/summary/50687/low-back-disorders?q=lumbar



58. National Institute for Health and Care Excellence. Percutaneous transforaminal

endoscopic lumbar discectomy for sciatica.

https://www.nice.org.uk/guidance/ipg556/resources/percutaneous-transforaminal-

endoscopic-lumbar-discectomy-for-sciatica-pdf-1899871990618309. Published 2016.

Accessed January 15, 2018.

59. National Institute for Health and Care Excellence. Percutaneous interlaminar endoscopic

lumbar discectomy for sciatica. https://www.nice.org.uk/guidance/ipg555. Published

2016. Updated April 27. Accessed November 8, 2017.

60. National Institute for Health and Care Excellence. Percutaneous coblation of the

intervertebral disc for low back pain and sciatica[IPG543]. Interventional procedures

guidance https://www.nice.org.uk/guidance/ipg543. Published 2016. Updated January

2016. Accessed January 15, 2018.

61. National Institute for Health and Care Excellence. Percutaneous electrothermal treatment

of the intervertebral disc annulus for low back pain and sciatica[IPG544]. Interventional

procedures guidance https://www.nice.org.uk/guidance/ipg544. Published 2016. Updated

January 2016. Accessed January 15, 2018, 2018.

62. National Institute for Health and Care Excellence. Percutaneous intradiscal

radiofrequency treatment of the intervertebral disc nucleus for low back pain[IPG545].

Interventional Procedures Guidance https://www.nice.org.uk/guidance/ipg545. Published

2016. Updated January 2016. Accessed January 15, 2018.

63. National Institute for Health and Care Excellence. Epiduroscopic lumbar discectomy

through the sacral hiatus for sciatica[IPG570]. Interventional procedures guidance

https://www.nice.org.uk/guidance/ipg570. Published 2016. Accessed January 15, 2018.

64. National Institute for Health and Care Excellence. Percutaneous intradiscal laser ablation

in the lumbar spine[IPG357]. Interventional procedures guidance

https://www.nice.org.uk/guidance/ipg357. Published 2010. Updated September 2010.

Accessed January 15, 2018.

65. National Institute for Health and Care Excellence. Automated percutaneous mechanical

lumbar discectomy. https://www.nice.org.uk/guidance/IPG141. Published 2005. Updated

November. Accessed November 8, 2017.

66. National Institute for Health and Care Excellence. Endoscopic laser

foraminoplasty[IPG31]. Interventional procedures guidance

https://www.nice.org.uk/guidance/ipg31. Published 2003. Accessed January 15, 2018.

67. Lewis R, Williams N, Matar HE, et al. The clinical effectiveness and cost-effectiveness

of management strategies for sciatica: systematic review and economic model. Health

Technol Assess. 2011;15(39):1-578. PMID: 22078311. doi: 10.3310/hta15390

68. Centers for Medicare & Medicaid Services. National Coverage Determination (NCD) for

Thermal Intradiscal Procedures (TIPs) (150.11). https://www.cms.gov/medicare-

coverage-database/details/ncd-

details.aspx?NCDId=324&ncdver=1&bc=AAAAgAAAAAAA&. Published n.d.

Accessed February 5, 2018.

69. Seikagaku Corporation. A Study of SI-6603 in Patients With Lumbar Disc Herniation.

Vol NCT019415632013.

70. Mayo Clinic. Epidural Clonidine for Lumbosacral Radiculopathy. Vol

NCT005883542008.

https://www.nice.org.uk/guidance/ipg556/resources/percutaneous-transforaminal-endoscopic-lumbar-discectomy-for-sciatica-pdf-1899871990618309

https://www.nice.org.uk/guidance/ipg556/resources/percutaneous-transforaminal-endoscopic-lumbar-discectomy-for-sciatica-pdf-1899871990618309

https://www.nice.org.uk/guidance/ipg555






https://www.nice.org.uk/guidance/IPG141


https://www.cms.gov/medicare-coverage-database/details/ncd-details.aspx?NCDId=324&ncdver=1&bc=AAAAgAAAAAAA&





71. Cook JA. The challenges faced in the design, conduct and analysis of surgical

randomised controlled trials. Trials. 2009;10:9-9. PMID: 19200379. doi: 10.1186/1745-

6215-10-9

72. Hsu PS, Armon C, Levin K. Acute lumosacral radiculopathy: Pathophysiology, clinical

features, and diagnosis. Waltham, MA: UpToDate Inc.; 2017. http://www.uptodate.com

73. Genevay S, Atlas SJ. Lumbar spinal stenosis. Best Pract Res Clin Rheumatol.

2010;24(2):253-265. PMID: 20227646. doi: 10.1016/j.berh.2009.11.001

74. Levin K, S. HP, C. A. Acute lumbosacral radiculopathy: Treatment and prognosis.

Waltham, MA: UpToDate Inc.; 2017. http://www.uptodate.com

75. Vroomen PC, de Krom MC, Knottnerus JA. Predicting the outcome of sciatica at short-

term follow-up. Br J Gen Pract. 2002;52(475):119-123. PMID: 11887877.

76. Wilson CA, Roffey DM, Chow D, Alkherayf F, Wai EK. A systematic review of

preoperative predictors for postoperative clinical outcomes following lumbar discectomy.

Spine J. 2016;16(11):1413-1422. PMID: 27497886. doi: 10.1016/j.spinee.2016.08.003

77. U.S. Food and Drug Administration. 510(k) Summary for Stryker DekompressorTM

Percutaneous Discectomy Probe.

https://www.accessdata.fda.gov/cdrh_docs/pdf3/K032473.pdf. Published 2003. Accessed

February 5, 2018.

78. U.S. Food and Drug Administration. Section 510(k) Premarket Notification for

Herniatome Percutaneous Discectomy Device.

https://www.accessdata.fda.gov/cdrh_docs/pdf14/k141557.pdf. Published 2014.


79. U.S. Food and Drug Administration. 510(k) Summary for Nucleotome®.

https://www.accessdata.fda.gov/cdrh_docs/pdf4/K040919.pdf. Published 2004. Accessed

February 5, 2018.

80. U.S. Food and Drug Administration. 510(k) Summary for Laurimed Percutaneous

Discectomy Sustem. https://www.accessdata.fda.gov/cdrh_docs/pdf8/k082194.pdf.

Published 2008. Accessed February 5, 2018.

81. U.S. Food and Drug Administration. Section 5: 510(k) Summary for ENSPIRETM

Debrider System. https://www.accessdata.fda.gov/cdrh_docs/pdf9/k090278.pdf.


82. U.S. Food and Drug Administration. 510(k) Summary for enSpireTM Discectomy System.

https://www.accessdata.fda.gov/cdrh_docs/pdf11/k110992.pdf. Published 2011.


83. U.S. Food and Drug Administration. 510(k) Summary of Safety and Effectivness for

Disc-FXTM System. https://www.accessdata.fda.gov/cdrh_docs/pdf5/K052241.pdf.


84. U.S. Food and Drug Administration. 510(k) Summary for ORATEC(R) Spine CATHTM

Intradiscal Catheter. https://www.accessdata.fda.gov/cdrh_docs/pdf/K993967.pdf.


85. U.S. Food and Drug Administration. 510(k) Summary of Safety and Effectiveness for

SpineCATH Intradiscal Catheter (Nucleotomy Catheter).

https://www.accessdata.fda.gov/cdrh_docs/pdf/K013622.pdf. Published 2002. Accessed

February 5, 2018.

86. U.S. Food and Drug Administration. 510(k) Summary for ArthroCare® Corporation,

ArthroCare® Coblator IQTM Perc-D® SpineWands®.

http://www.uptodate.com/

http://www.uptodate.com/


https://www.accessdata.fda.gov/cdrh_docs/pdf14/k141557.pdf






https://www.accessdata.fda.gov/cdrh_docs/pdf/K993967.pdf




https://www.accessdata.fda.gov/cdrh_docs/pdf13/K132099.pdf. Published 2013.


87. U.S. Food and Drug Administration. 510(k) Premarket Notification for METRX

SYSTEM from Richard W. Treharne, Medtronic Sofamore Danek, Inc.

https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K002931.

Published 2018. Updated January 29. Accessed February 5, 2018.

88. U.S. Food and Drug Administration. 510(k) Summary of Safety and Effectivess for

Yeung Endoscopic Spine System.

https://www.accessdata.fda.gov/cdrh_docs/pdf/K973405.pdf. Published 1998. Accessed

February 5, 2018.

89. North American Spine Society. Minimally Invasive Procedures for Back and Neck Pain.

https://northamericanspine.com/. Published n.d. Accessed February 5, 2018.

90. U.S. Food and Drug Administration. 510(k) Summary for Ceralas Diode 980nm Laser

System (Models D15, D25).

https://www.accessdata.fda.gov/cdrh_docs/pdf7/K072779.pdf Published 2008. Accessed

February 5, 2018.

91. Washington State Health Care Authority. Surgery for Symptoms of Lumbar

Radiculopathy: Daft key questions: public comment and response.

https://www.hca.wa.gov/about-hca/health-technology-assessment/surgery-for-

symptomatic-lumbar-radiculopathy Published 2017. Updated November 17. Accessed

February 3, 2018.

92. Hallgren KA. Computing inter-rater reliability for observational data: an overview and

tutorial. Tutor Quant Methods Psychol. 2012;8(1):23-34. PMID: 22833776.

93. Gartlehner G, West SL, Mansfield AJ, et al. Clinical heterogeneity in systematic reviews

and health technology assessments: synthesis of guidance documents and the literature.

Int J Technol Assess Health Care. 2012;28(1):36-43. PMID: 22217016. doi:

10.1017/s0266462311000687

94. West SL, Gartlehner G, Mansfield AJ, et al. AHRQ Methods for Effective Health Care.

Comparative Effectiveness Review Methods: Clinical Heterogeneity. Rockville (MD):

Agency for Healthcare Research and Quality (US); 2010.

95. Peul WC, van den Hout WB, Brand R, Thomeer RT, Koes BW. Prolonged conservative

care versus early surgery in patients with sciatica caused by lumbar disc herniation: two

year results of a randomised controlled trial. BMJ. 2008;336(7657):1355-1358. PMID:

18502911. doi: 10.1136/bmj.a143

96. Lequin MB, Verbaan D, Jacobs WC, et al. Surgery versus prolonged conservative

treatment for sciatica: 5-year results of a randomised controlled trial. BMJ Open.

2013;3(5). PMID: 23793663. doi: 10.1136/bmjopen-2012-002534

97. Brouwer PA, Brand R, van den Akker-van Marle ME, et al. Percutaneous laser disc

decompression versus conventional microdiscectomy for patients with sciatica: Two-year

results of a randomised controlled trial. Interv Neuroradiol. 2017;23(3):313-324. PMID:

28454511. doi: 10.1177/1591019917699981

98. Gempt J, Jonek M, Ringel F, Preuss A, Wolf P, Ryang Y. Long-term follow-up of

standard microdiscectomy versus minimal access surgery for lumbar disc herniations.

Acta Neurochir (Wien). 2013;155(12):2333-2338. PMID: 24136677. doi:

10.1007/s00701-013-1901-z


https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K002931


https://northamericanspine.com/


https://www.hca.wa.gov/about-hca/health-technology-assessment/surgery-for-symptomatic-lumbar-radiculopathy

https://www.hca.wa.gov/about-hca/health-technology-assessment/surgery-for-symptomatic-lumbar-radiculopathy



99. Barth M, Weiss C, Thome C. Two-year outcome after lumbar microdiscectomy versus

microscopic sequestrectomy: part 1: evaluation of clinical outcome. Spine (Phila Pa

1976). 2008;33(3):265-272. PMID: 18303458. doi: 10.1097/BRS.0b013e318162018c

100. Hawker GA, Mian S, Kendzerska T, French M. Measures of adult pain: Visual Analog

Scale for Pain (VAS Pain), Numeric Rating Scale for Pain (NRS Pain), McGill Pain

Questionnaire (MPQ), Short-Form McGill Pain Questionnaire (SF-MPQ), Chronic Pain

Grade Scale (CPGS), Short Form-36 Bodily Pain Scale (SF-36 BPS), and Measure of

Intermittent and Constant Osteoarthritis Pain (ICOAP). Arthritis Care Res (Hoboken).

2011;63(S11):S240-252. PMID: 22588748. doi: 10.1002/acr.20543

101. Ward MM, Guthrie LC, Alba MI. Clinically important changes in short form-36 scales

for use in rheumatoid arthritis clinical trials: the impact of low responsiveness. Arthritis

care & research. 2014;66(12):1783-1789. PMID: PMC4245332. doi: 10.1002/acr.22392

102. Maruish ME. User's manual for the SF-36v2 Health Survey. Lincoln, RI: QualityMetric

Inc; 2011.

103. Schwind J, Learman K, O'Halloran B, Showalter C, Cook C. Different minimally

important clinical difference (MCID) scores lead to different clinical prediction rules for

the Oswestry disability index for the same sample of patients. J Man Manip Ther.

2013;21(2):71-78. PMID: 24421616. doi: 10.1179/2042618613y.0000000028

104. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine (Phila Pa 1976).

2000;25(22):2940-2952; discussion 2952. PMID: 11074683.

105. Fu YS, Zeng BF, Xu JG. Long-term outcomes of two different decompressive techniques

for lumbar spinal stenosis. Spine (Phila Pa 1976). 2008;33(5):514-518. PMID:

18317196. doi: 10.1097/BRS.0b013e3181657dde

106. Javid MJ, Nordby EJ, Ford LT, et al. Safety and efficacy of chymopapain

(Chymodiactin) in herniated nucleus pulposus with sciatica. Results of a randomized,

double-blind study. JAMA. 1983;249(18):2489-2494. PMID: 6341632.

107. J. AS, Singer DE, L. PD, Deyo RA. The efficacy of lumbar disc surgery in patients with

sciatica: preliminary results of the Maine Lumbar Spine Study. Clin Res. 1993;41:200A.

PMID.

108. Premera Blue Cross. Medical Policy - 7.01.93. Decompression of the Intervertebral Disc

Using Laser Energy (Laser Discectomy) or Radiofrequency Coblation (Nucleoplasty).

https://www.premera.com/medicalpolicies/7.01.93.pdf. Published 2017. Accessed

February 5, 2018.

109. Premera Blue Cross. Medical Policy - 7.01.551. Lumbar Spine Decompression Surgery:

Discectomy, Foraminotomy, Laminotomy, Laminectomy.

https://www.premera.com/medicalpolicies/7.01.551.pdf. Published 2017. Accessed

February 5, 2018.

110. Premera Blue Cross. Medical Policy - 7.01.18. Automated Percutaneous and

Percutaneous Endoscopic Discectomy.

https://www.premera.com/medicalpolicies/7.01.18.pdf Published 2017. Accessed

February 5, 2018.

111. Regence BlueCross, BlueShield of Oregon. Surgery, Policy No. 145. Automated

Percutaneous and Percutaneous Endoscopic Discectomy.

http://blue.regence.com/trgmedpol/surgery/sur145.pdf. Published 2017. Accessed

February 5, 2018.

https://www.premera.com/medicalpolicies/7.01.93.pdf



http://blue.regence.com/trgmedpol/surgery/sur145.pdf



112. Regence BlueCross, BlueShield of Oregon. Surgery, Policy No. 131. Decompression of

Intervertebral Discs Using Laser Energy (Laser Discectomy) or Radiofrequency Energy

(Nucleoplasty). http://blue.regence.com/trgmedpol/surgery/sur131.pdf. Published 2017.


113. UnitedHealthcare. Discogenic Pain Treatment.

https://www.uhcprovider.com/content/dam/provider/docs/public/policies/comm-medical-

drug/discogenic-pain-treatment.pdf Published 2017. Accessed February 5, 2018.

114. Cigna. Minimally Invasive Intradiscal/ Annular Procedures and Trigger Point Injections.

https://cignaforhcp.cigna.com/public/content/pdf/coveragePolicies/medical/mm_0139_co

veragepositioncriteria_invasive_treatment_for_back_pain.pdf. Published 2017. Accessed

February 5, 2018.

115. Humana. Spinal Decompression Surgery.

http://apps.humana.com/tad/tad_new/Search.aspx?criteria=laminectomy&searchtype=fre

etext&policyType=both. Published 2017. Accessed February 5, 2018.

116. NC Division of Medical Assistance. Medicaid and Health Choice. Spinal Surgeries,

Clinical Coverage Policy No: 1A-30. https://files.nc.gov/ncdma/documents/files/1A-

30.pdf. Published 2017. Accessed February 5, 2018.

http://blue.regence.com/trgmedpol/surgery/sur131.pdf

https://www.uhcprovider.com/content/dam/provider/docs/public/policies/comm-medical-drug/discogenic-pain-treatment.pdf

https://www.uhcprovider.com/content/dam/provider/docs/public/policies/comm-medical-drug/discogenic-pain-treatment.pdf

https://cignaforhcp.cigna.com/public/content/pdf/coveragePolicies/medical/mm_0139_coveragepositioncriteria_invasive_treatment_for_back_pain.pdf

https://cignaforhcp.cigna.com/public/content/pdf/coveragePolicies/medical/mm_0139_coveragepositioncriteria_invasive_treatment_for_back_pain.pdf

http://apps.humana.com/tad/tad_new/Search.aspx?criteria=laminectomy&searchtype=freetext&policyType=both

http://apps.humana.com/tad/tad_new/Search.aspx?criteria=laminectomy&searchtype=freetext&policyType=both

https://files.nc.gov/ncdma/documents/files/1A-30.pdf

https://files.nc.gov/ncdma/documents/files/1A-30.pdf


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page A-1

Appendix A. Search Strategy

PubMed searched from 1/1/2007-11/09/2017

((((((((((((((((((((((((((((((((((((laser disc decompression[Title/Abstract]) OR laser thermo-

discoplasty[Title/Abstract]) OR laser discectomy[Title/Abstract]) OR laser disc

thermoplasty[Title/Abstract]) OR laser disc thermoplasty[Title/Abstract]) OR ((laser

thermoplasty[Title/Abstract] AND disc[Title/Abstract]))) OR laser thermoplasty[Title/Abstract]) OR laser

nucleotomy[Title/Abstract]) OR MILD[Title/Abstract]) OR minimally invasive lumbar

decompression[Title/Abstract]) OR image-guided lumbar decompression[Title/Abstract]) OR thermal

discoplasty[Title/Abstract]) OR ((endoscopic discectomy[Title/Abstract]) OR endoscopic

nucleotomy[Title/Abstract])) OR ((percutaneous nucleotomy[Title/Abstract]) OR percutaneous

discectomy)) OR thermal nucleoplasty[Title/Abstract]) OR laser discoplasty[Title/Abstract])))) AND

Lumbar[Title/Abstract])) OR (((((((((((((((((((((((Laminectomy[MeSH Terms]) OR Laminotomy[MeSH

Terms]) OR Diskectomy[MeSH Terms]) OR Diskectomy, percutaneous[MeSH Terms]) OR

Foraminotomy[MeSH Terms]) OR Intervertebral Disc/surgery[MeSH Terms]) OR Intervertebral

Disc/therapy[MeSH Terms]) OR Intervertebral Disc Displacement/surgery[MeSH Terms]) OR

Intervertebral Disc Displacement/therapy[MeSH Terms]) OR Intervertebral Disc

Degeneration/surgery[MeSH Terms]) OR Intervertebral Disc Degeneration/therapy[MeSH Terms]) OR

Sciatica/surgery[MeSH Terms]) OR Sciatica/therapy[MeSH Terms]) OR Spinal

osteophytosis/surgery[MeSH Terms]) OR Spinal Nerve Roots/surgery[MeSH Terms]) OR Spinal

osteophytosis/therapy[MeSH Terms]) OR Lumbar vertebrae/surgery[MeSH Terms]) OR Lumbar

vertebrae/therapy[MeSH Terms])))))))) NOT (Goats[MeSH Terms] OR Mouse[MeSH Terms] OR

Rat[MeSH Terms] OR Dogs[MeSH]))) NOT (Cervical vertebrae[MeSH Terms] OR Cervical

vertebrae[Title/Abstract]))) NOT (Cancer[MeSH Terms] OR Fractures[MeSH Terms] OR Cancer

Pain[MeSH Terms] OR Tuberculosis[MeSH Terms] OR Spinal Cord Injuries[MeSH Terms] OR

Scoliosis[MeSH Terms]))) NOT (Case Reports[Publication Type] OR Comment[Publication Type] OR

Letter[Publication Type] OR Patient Education Handout[Publication Type] OR Editorial[Publication

Type]))) NOT (children[MeSH Terms] OR infants[MeSH Terms] OR adolescents[MeSH Terms]))) NOT

Spinal Fusion[MeSH Major Topic]) NOT Spondylolisthesis/surgery[MeSH Major Topic])) AND

(Comparative Study[Publication Type] OR Review[Publication Type] OR Meta-analysis[Publication Type]

OR Clinical Study[Publication Type] OR Randomized Controlled Trial[Publication Type] OR Clinical

Trial[Publication Type] OR Clinical Trial, Phase I[Publication Type] OR Clinical Trial, Phase II[Publication

Type] OR Clinical Trial, Phase III[Publication Type] OR Clinical Trial, Phase IV[Publication Type] OR

Observational Study[Publication Type] OR Pragmatic Clinical Trial[Publication Type] OR Controlled

Clinical Trial[Publication Type]) Sort by: PublicationDate Filters: Publication date from 2007/01/01 to

2017/11/09; English

Yield: 1,443

Cochrane Library Search from inception to 11/10/2017

Terms: Sciatica, lumbar radiculopathy, lumbar disc disease, minimally invasive spine

Total Yield: 93


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page A-2

ClinicalTrials.Gov Search from inception to 11/10/2017

Terms: Sciatica, lumbar radiculopathy, lumbar disc disease, minimally invasive spine; limits: Adult 18-

65, Adult 66+, Interventions

Total Yield: 253

Other Data

The following websites were searched using the terms radiculopathy, laminectomy, discectomy,

practice guidelines, spine surgery, nerve root compression

U.S. Food and Drug Administration

Centers for Medicare and Medicaid Services

Aetna

UnitedHealth

Humana

BlueCross BlueShield (Premera and Regence)

Kaiser Permanente

National Institute for Health and Care Excellence (U.K.)

U.S. Agency for Healthcare Research and Quality

North American Spine Society

American Society of Interventional Pain Physicians

American Academy of Orthopaedic Surgeons

American Academy of Neurological Surgeons

American Pain Society

American College of Occupational and Environmental Medicine


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page B-1

Appendix B. Additional Methods

The following exchanges rates were used to convert foreign costs reported to U.S. dollars:

U.S. $ British Pound Euro €

Year 1992 1 0.568 -

Year 2008 1 - 0.659

Year 2009 1 - 0.789

Year 2010 1 - 0.740

Source: U.S. Department of Treasury. Treasury Reporting Rates of Exchange. Historical Rates for March 31st, 1992; March 31st,

2008; March 31st, 2009; and March 31st, 2010. Available at:

https://www.fiscal.treasury.gov/fsreports/rpt/treasRptRateExch/historicalRates.htm Accessed January 21, 2018.

The following chain-weighted, average year consumer price indices were used to adjust all

reported costs to 2010 dollars.

Year Annual Average CPI

1992 140.3

1993 144.5

2004 188.9

2008 215.30

2009 214.54

2010 218.06

Source: U.S. Department of Labor, Bureau of Labor Statistics. CPI Databases. All Urban Consumers (Chained CPI). Average

Annual Indices. Available at: https://www.bls.gov/cpi/data.htm . Accessed January 21, 2018.

https://www.fiscal.treasury.gov/fsreports/rpt/treasRptRateExch/historicalRates.htm

https://www.bls.gov/cpi/data.htm


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page C-1

Appendix C. Evidence Tables

Table C-1. Study characteristics of included studies



Table C-2. Population characteristics of included studies

Table C-3. Individual study findings related to efficacy outcomes Part 1 (pain, neurological symptoms, quality of life)

Table C-4. Individual study findings related to efficacy outcomes Part 2 (functioning/disability, return to work, other efficacy outcomes)

Table C-5. Individual study findings related to safety outcomes

Table C-6. Individual study findings related to cost outcomes



Table C-1. Study characteristics of included studies

Author (Year)

Study Design; Country; Risk of Bias

Study Sponsor; Study Sponsor Name; Trial Name (if applicable)

Surgical Intervention; N randomized; N analyzed (% of randomized); N crossovers (% of randomized); Intervention Description

Comparator(s); N randomized; N analyzed (% of randomized); N crossovers (% of randomized); Intervention Description

Arts (2009)38 Arts (2011)43

Parallel-group RCT; The Netherlands; Low

Government; The Dutch Health Care Insurance Board; The Sciatica Micro-Endoscopic Discectomy Randomized Controlled Trial

Tubular discectomy N randomized: 167; N analyzed: 166 (99.4%); N crossovers: 2 (1.2%); Surgery scheduled within 4w of first visit. A 25-30 mm midline incision used, skin retracted laterally and the guidewire and sequential dilators (METRx, Medtronic, Minneapolis, Minnesota) were placed at the inferior aspect of the lamina under fluoroscopic control. A 14- to 18-mm working channel was introduced over the final dilator and attached to the table. The herniated disk was removed through the tubular retractor with microscopic magnification. Bony lamina removal was minimal, if necessary.

Microdiscectomy N randomized: 161; N analyzed: 159 (98.8%); N crossovers: 0 (0%); Surgery was scheduled within 4w of first visit. A 25-30mm midline skin incision used followed by ipsilateral paravertebral muscle retraction. The herniated disk was removed by the unilateral transflaval approach with the aid of a headlight loupe or microscope magnification, depending on the surgeon’s preference. Bony lamina removal was minimal, if necessary.

Brouwer (2015)37 Brouwer (2017) 97

Parallel-group RCT; The Netherlands; Some concerns

Government; Healthcare Insurance Board of the Netherlands

Percutaneous laser disc decompression N randomized: 57; N analyzed: 55 (96.5%); N crossovers: Unclear; CT–guided treatment was performed with the patient in prone position under local anesthesia. An 18-G needle was placed centrally in the nucleus pulposus and parallel to the end plates by means of a posterolateral approach. Through the needle, a glass fiber of 600 micron was advanced into the disc, enabling the application of laser energy (diode laser; Biolitec Inc, East Longmeadow, MA, USA; 980 nm, 7 W, 0.6-second pulses, and an interval of 1 second) for a total energy delivered of 1,500 J (2,000 J for L4–L5 level).

Microdiscectomy N randomized: 58; N analyzed: 57 (98.3%); N crossovers: 0 (0%); A discectomy performed under general or spinal anesthesia using loupe magnification or microscope depending on the surgeon’s preference. The aim of the surgery was to remove the herniated disc fragment, without any attempt to remove the disc itself, using a unilateral transflaval approach.

(continued)



Table C-1. Study characteristics of included studies (continued)

Author (Year)





Chatterjee (1995)36

Parallel-group RCT; United Kingdom; Some concerns

Government; The Department of Health, London, UK

Automated Percutaneous Lumbar Discectomy N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%); Procedure was performed with a 2-mm nonflexible automated suction nucleotome (Surgical Dynamics, San Leandro, California) under local anesthesia and with biplanar radiologic control. It was necessary to achieve a position that was either exactly central within the disc or slightly posterior to center before disc aspiration was commenced. Disc aspiration was continued until no more nuclear material could be obtained.

Microdiscectomy N randomized: 40; N analyzed: 40 (100%); N crossovers: 0 (0%); Microdiscectomy was performed by standard technique via a 2-cm incision and a transligamentous approach with the removal of not only the herniated portion of the disc but also with clearance of all loose intradiscal material.

Erginousakis (2011)35

Parallel-group RCT; Greece; High

Reported as NOT industry supported.

Percutaneous disc decompression N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%) in main analysis; 20 received subsequent microdiscectomy Intervertebral disc decompression using a 17-gauge Dekompressor (Stryker, Kalamazoo, MI) was performed with fluoroscopic guidance. Approximately 1–3 grams of disc material and 1 milliliter of tissue has been removed once the tissue becomes visible at the collection chamber entrance.

Conservative Management N randomized: 31; N analyzed: 31 (100%); N crossovers: 0 (0%); A 6-week course of monitored and registered conservative therapy during which participants received analgesics, anti-inflammatory drugs, muscle relaxants, and physiotherapy. It also included education and counseling.

(continued)




Author (Year)





Franke (2009)34

Parallel-group RCT; Germany; Some concerns

NR

Microscopically assisted percutaneous nucleotomy N randomized: 52; N analyzed: 52 (100%); N crossovers: 0 (0%); The level localization with a spinal needle was done on the opposite side. The pinpoint was directed at the open interlaminar window. The skin incision of 15 mm at the side of the pathology was performed in height of the needle entry point approximately 2 cm paramedian. Both the thoracolumbar fascia and the paraspinal muscles were dilated till the working channel could be brought in; surgery performed under direct vision via a microscope

Microdiscectomy N randomized: 48; N analyzed: 48 (100%); N crossovers: 0 (0%); Procedure not described.

Gerszten (2010)39

Parallel-group RCT; United States; Some concerns (6w outcomes) High (12w and later outcomes)

Commercial; ArthroCare Corp.

Plasma disc decompression with coblation technology (PDD) N randomized: 46; 45 ITT sample; N analyzed: 29 (64% of ITT sample) at 26w; N crossovers: 12 were unresolved and received a second, unspecified procedure; Procedure performed on an outpatient basis using the Coblation DLR or DLG SpineWand surgical device (ArthroCare Corp.). Procedure conducted under fluoroscopic guidance. A 17G spinal cannula was introduced into disc using a posterolateral extrapedicular approach and position at the junction of the annulus and nucleus. The SpineWand was introduced through the cannula and positioned within the nucleus then placed in ablation mode and advanced and retracted to create a total of 6 channels.

Epidural steroid injection (ESI) N randomized: 44; 40 ITT sample; N analyzed: 28 (70% of ITT sample) at 26w; N crossovers: 8 were unresolved and received a second, unspecified procedure; Procedure performed under fluoroscopic guidance. The location was determined by the treating physician with the goal of delivering steroids to the site of the disc protrusion and nerve irritation. A transforaminal approach was used. Medication type and dose were left to the discretion of the treating clinician. A second ESI procedure was allowed by the study protocol.

(continued)




Author (Year)





Haines (2002)40 Parallel-group RCT; United States; High

Government; National Institute of Neurological Disorders and Stroke and the Agency for Health Care Research and Quality; LAPDOG

Automated percutaneous discectomy, endoscopic percutaneous discectomy (APD/EPD) N randomized: 21; N analyzed: 17 (81.0%) at 26w; N crossovers: 0 (0%); When this study was conceived there was a single manufacturer of the only device specifically designed for APD (Nucleotome1, Surgical Dynamics, Alameda, CA, USA). Surgical Dynamics had produced an explicit protocol specifying inclusion and exclusion criteria as well as surgical technique. This was adhered to in the APD group. With technology advancement and diffusion, equipment for endoscopic approach became available from several manufacturers and indications for the procedure became less specific. Thus, the intervention protocol was modified after enrollment of the first 26 participants to incorporate epidural endoscopic technique.

Discectomy N randomized: 13; N analyzed: 10 (76.9%) at 26w; N crossovers: 1 (7.69%); Discectomies were done according to the surgeon's usual technique.

Henriksen (1996)33

Parallel-group RCT; Denmark; Some concerns

NR

Microdiscectomy N randomized: 40; N analyzed: 39 (97.5%); N crossovers: 0 (0%); The patients were operated in the prone position. A midline skin incision approximately 7 cm long was used over the appropriate disc, and a 7 cm long fascial incision was made. Additional details provided but not clear whether they pertain to SG1 or SG2.

Discectomy N randomized: 40; N analyzed: 40 (100%); N crossovers: 0 (0%); The patients were operated in the prone position. A midline skin incision approximately 7 cm long was used over the appropriate disc, and a 3 cm long fascial incision was made. Additional details provided but not clear whether they pertain to SG1 or SG2. For example, 1/3 of the facet joint was removed. An editorial note associated with the study raised concerns about whether the procedure performed in SG2 was what most would consider a microsurgical discectomy.

(continued)




Author (Year)





Hermantin (1999)41

Parallel-group RCT; United States; Some concerns

NR

Video-assisted arthroscopic microdiscectomy N randomized 30; N analyzed 30 (100%); N crossovers 0 (0%); Performed with the use of an oval 5X8mm cannula introduced through a universal cannula by an 18G needle under fluoroscopy, positioned in the triangular working zone bordered anterolaterally by the exiting nerve root, medially by the traversing nerve root and dura, and caudally by the vertebral plate of the caudad lumbar segment. Herniated disc fragments are pulled back into the intervertebral disc space and then are withdrawn.

Discectomy, with laminotomy N randomized: 30; N analyzed: 30 (100%); N crossovers: 0 (0%); The open laminotomy and discectomy was performed in a standard fashion. A four-centimeter posterior midline incision was made, and a small laminotomy and discectomy was performed at the specified level.

Huang (2005)23

Parallel-group RCT; Taiwan; Some concerns

Other; Research Committee of Chang Gung Memorial Hospital, Taiwan

Microendoscopic discectomy N randomized: 10; N analyzed: 10 (100%); N crossovers: 0 (0%); The Vertebroscope System (Zeppelin, Pullach, Germany) was used to perform the endoscopic discectomy procedure.

Discectomy N randomized: 12; N analyzed: 12 (100%); N crossovers: 0 (0%); Intervention not described.

Malter (1996)42 Cost-effectiveness Analysis; United States;

Government; Agency for Health Care Policy and Research, Seattle VA Medical Center, National Research Service Award

Discectomy N randomized: NA

Conservative management N randomized: NA

(continued)




Author (Year)





Mayer (1993)32 Parallel-group RCT; Germany; High

NR

Percutaneous endoscopic discectomy N randomized: 20; N analyzed: 20 (100%); N crossovers: 3 (15%); The procedure is performed with the patient under local anesthesia, using approach is as described by Day and Nazarian. Under fluoroscopic control, the tip of a 18G cannula is advanced to the center of the disc. Discography is performed to confirm the indication for percutaneous endoscopic discectomy. A guidewire is advanced through the cannula until the tip reaches the center of the disc; the cannula is then removed. With the wire as a guide, a blunt tapered trocar is advanced to the posterolateral border of the annulus fibrosus through a stab incision. The trocar, in turn, serves as a guide for the introduction of the working cannula (outer diameter 5 mm). Following introduction of the working cannula, the trocar is removed and the disc is entered by cutting a circular window in the annulus fibrosus with the aid of a trephine. Rigid forceps are introduced to remove a small amount of nucleus pulposus from the center of the disc in order to create a cavity before introducing the endoscope. Endoscopy of the disc is performed with a rigid endoscope. The herniated part of the nucleus pulposus can be removed using reverse-opening forceps as well as flexible forceps. A bilateral approach is used for continuous endoscopy during removal of disc herniations located in the midline.

Microdiscectomy N randomized: 20; N analyzed: 20 (100%); N crossovers: 0 (0%); Intervention not described.

(continued)




Author (Year)





McMorland (2010)22

Parallel-group RCT; Canada; Some concerns

Other; Supported by a grant from the Foundation for Chiropractic Education and Research.

Microdiscectomy N randomized: 20; N analyzed: 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w 20 (100%) 52w 15 (75%); N crossovers: 3 (15%) enrolled in spinal manipulation 26-34w after surgery so received both interventions. Note, this crossover happened AFTER all ITT outcomes are reported (12w); Surgical microdiscectomies were performed with patients in a prone position supported by bolsters through a standard midline lumbar incision. All procedures were undertaken using microsurgical techniques with the aid of an operating microscope. Laminotomies were created as required at the level of the lumbar disc herniation. Both sequestrectomy and intra-annular discectomy were performed to ensure adequate nerve root decompression.

Spinal manipulation N randomized: 20; N analyzed: 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w 20 (100%) 52w 17 (85%); N crossovers: 8 (40%) underwent microdiscectomy after 12w of spinal manipulation care. Note, this crossover happened AFTER all ITT outcomes are reported (12w); All spinal manipulative therapies were provided by a single chiropractic doctor. Spinal manipulative therapy consisted of side posture, high-velocity, low-amplitude, short lever technique. The decision to administer manual spinal manipulation on each visit was based on that patient's ability to tolerate the position. Cryotherapy or thermotherapy (ice or heat) were used on an “as-needed” basis. All patients were provided with an information/education package and were introduced to rehabilitative exercises. The patients also participated in a supervised rehabilitative (core stability) exercise regimen. Treatments typically required 2 to 3 visits per week for the first 4 weeks reducing to 1 to 2 visits per week for the next 3 to 4 weeks. At the 8-week mark, follow-up visits were scheduled based on the patient's symptoms until the patient's symptoms were deemed stable (i.e., no deterioration or flare up) with a 2-month treatment holiday. Mean number of treatment session was 21 plus an additional 6 supervised rehabilitation sessions over 52 weeks.

(continued)




Author (Year)





North (2005)47 Parallel-group RCT; United States; High

Commercial; Medtronic, Inc.

Repeat lumbosacral decompression N randomized: 26; N analyzed: 26 (100%); N crossovers: 14 (54%); Laminectomy (N=23) and/or foraminotomy (N=21) and/or discectomy(N=6) with or without fusion (N=3), with or without instrumentation (N=6). Patients randomized to reoperation could cross over to spinal cord stimulation after a 6-month postoperative period.

Spinal cord stimulation N randomized: 24; N analyzed: 19 (79.2%); N crossovers: 5 (20.8%); Percutaneous placement of a temporary electrode (3487A Pisces-Quad; Medtronic, Inc., Minneapolis, Minnesota) for a therapeutic trial lasting at least 3 days. The spinal cord stimulation patients could receive a permanent implant (3487A-56 or 3587A Resume electrode, X-trel or Itrel pulse generator; Medtronic, Inc.) if they reported at least 50% estimated relief of pain by standard pain rating methods and demonstrated stable or improved analgesic medication intake, with improved physical activity commensurate with neurological status and age. Patients randomized to spinal cord stimulation who did not meet these criteria could immediately cross over to reoperation.

(continued)




Author (Year)





Osterman (2003)31

Parallel-group RCT; Finland; High

Government; Finnish Office for Health Technology Assessment at National Research and Development Centre for Welfare and Health, Jorvi Hospital, Helsinki and Uusimaa Hospital District, Espoo, Finland.

Microdiscectomy N randomized: 28; N analyzed 6w 26 (93%) 12w 26 (93%) 26w 26 (93%) 52w 21 (75%) 2y 26 (93%); N crossovers: 0 (0%); A microdiscectomy by a spinal orthopedic surgeon was performed within 2 weeks of randomization. The operation was carried out under general anesthesia in a genupectoral position with fluoroscopic control of the spinal level before draping. The patients were usually discharged from the hospital on the second or third postoperative day. Sick leave and analgesia were prescribed according to individual requirements. Surgical patients were advised to continue with isometric exercises while waiting for the operation and also after discharge from the hospital. At follow-up visits, this group received active physiotherapeutic instructions, including stretching, bending, and muscle strengthening exercises. Passive forms of treatment were not recommended.

Physiotherapy N randomized: 28; N analyzed 6w 26 (93%) 12w 26 (93%) 26w 22 (78.6%) 52w 20 (71.4%) 2y 24 (86%); N crossovers” 11 (39.3%) Note: 3 were < 6w, 4 were between 6w and 12 w, 3 were between 12w and 26w, and 1 between 26w and 52w; The control group received physiotherapeutic instructions initially and continued with isometric exercises after randomization. At follow-up visits, as with the surgical group, activity was encouraged. Patients in the control group were informed of symptoms meriting operation, and they were advised to contact the treating physician if the symptoms should get worse.

(continued)




Author (Year)





Peul (2007)30 Peul (2008)95 Lequin (2013)96

Parallel-group RCT; The Netherlands; High

Government; Supported by a grant from the Netherlands Organisation for Health Research and Development (ZonMW) and the Hoelen Foundation, The Hague; Sciatica Trial

Microdiscectomy N randomized: 141; N analyzed 52w: 140 (99.3%) 2y: 130 (92.2%) 5y: 115 (81.6%); N crossovers 52w: 16 (11.3%) 2y: 16 (11.3%) 5y: 16 (11.3%); Surgery was scheduled within 2 weeks after assignment. A minimal unilateral transflaval approach with magnification. The goal was to decompress the nerve root and reduce the risk of recurrent disk herniation by performing an annular fenestration, curettage, and removal of loose degenerated disk material from the disk space with the use of a rongeur, without attempting to perform a subtotal discectomy.

Conservative management N randomized: 142; N analyzed 52w: 141 (99.3%) 2y: 130 (91.5%) 5y: 116 (81.7%); N crossovers 52w: 55 (38.7%) 2y: 62 (43.7%) 5y: 66 (46.5%); General practitioners provided prolonged conservative treatment. Invitation to website for education on natural course of their illness and the expectation of successful recovery, irrespective of the initial intensity of their pain. Treatment aimed at enabling resumption of daily activities. Pain medication as needed. Patients who were fearful of moving were referred to a physiotherapist. If sciatica persisted for 6 months after the patient underwent randomization, microdiscectomy was offered. Patients who had increasing leg pain not responsive to medication or progressive neurologic deficits were offered surgery earlier than 6 months after randomization.

(continued)




Author (Year)





Ruetten (2008)29 Parallel-group RCT; Germany; High

Other; Study reports that no funds or benefits were received or will be received in support of this work.

Endoscopic (interlaminar or transforaminal) discectomy N randomized: 100; N analyzed: 91 (91%); N crossovers: 0 (0%); All of the operating instruments and optics were products supplied by WOLF (Richard Wolf GmbH, Knittlingen, Germany). The full-endoscopic transforaminal procedure was used for extraforaminal and intraforaminal herniations. The full-endoscopic interlaminar operation was performed for herniations mainly in the spinal canal. Sequestrotomy alone was performed in small or covered annular defects when the sequestered disc material exceeded the level of the intervertebral space toward cranial or caudal. This occurred in 39 participants.

Microdiscectomy N randomized: 100; N analyzed: 87 (87%); N crossovers: 0 (0%); The conventional discectomy was performed with paramedian or lateral access in known standardized technique using a microscope. Sequestrotomy alone was performed in small or covered annular defects when the sequestered disc material exceeded the level of the intervertebral space toward cranial or caudal. This occurred in 43 participants.

(continued)




Author (Year)





Ruetten (2009)48 Parallel-group RCT; Germany; High

NR

Revision endoscopic discectomy N randomized: 50; N analyzed: 45 (90%); N crossovers: 0 (0%); Used either the transforaminal (TF) or interlaminar (IL) approach. Osseous resection was required in 6% of cases. The TF procedure was performed with access as lateral as possible. A spinal cannula is inserted via the 6mm skin incision. After insertion of a lead wire, the cannulated dilator is pushed in and a surgical sheath is placed. Decompression is performed. If the anatomic osseous diameter of the intervertebral foramen does not permit direct entry into the spinal canal, the opening is expanded. An extraforaminal approach is made at the caudal pedicle in cases where the position of the exiting nerve is not clear. The IL operation was performed using a dilator inserted bluntly to the lateral edge of the interlaminar window and an operation sheath directed toward the ligamentum flavum. The medial edge of the descending facet is located and prepared directly on the bone toward ventral until the medial edge of the ascending facet is visible. Blunt penetration to the floor of the spinal canal and preparation of the ventral epidural space. Bone resection to expand the interlaminar window to enable penetration into the spinal canal with the endoscope is usually not necessary owing to the resection during the primary operation. Sequestrotomy alone was performed in small or covered annular defects when the sequestered disc material has exceeded the level of the intervertebral space. All the operating instruments and optics were products supplied by WOLF (Richard Wolf GmbH, Knittlingen, Germany).

Revision microdiscectomy N randomized: 50; N analyzed: 42 (84%); N crossovers: 0 (0%); The conventional microsurgical operations were performed with paramedian access in known standardized technique using a microscope. Sequestrotomy alone was performed in small or covered annular defects when the sequestered disc material has exceeded the level of the intervertebral space toward cranial or caudal. Osseous resection was required in 94% of cases.

(continued)




Author (Year)





Ryang (2008)28 Gempt (2013)98

Parallel-group RCT; Germany; High

NR

Trocar microdiscectomy N randomized: 30; N analyzed: unclear; N crossovers: 0 (0%); The skin was incised horizontally over a length of 4 to 5 cm on the affected side after localization of the interlaminar space with lateral x-ray fluoroscopy. The lumbodorsal fascia was incised vertically over a distance of 4 to 5 cm, 0.5 cm paramedially. The paraspinal musculature was partially detached from the hemilamina in a subperiostal fashion, the interlaminar space was visualized, and the retractor placed into position. A second fluoroscopy was obtained to confirm the correct level. The operating microscope (Carl Zeiss Co., Oberkochen, Germany) was put into position and the remaining operation performed in the standard microsurgical fashion with bayoneted microsurgical instruments. Partial hemilaminectomy of the superior and inferior lamina and medial facetectomy, with partial flavectomy were carried out to visualize the compromised nerve root. The herniated sequester was removed. A partial nucleotomy was performed in some cases.

Microdiscectomy N randomized: 30; N analyzed: unclear; N crossovers: 0 (0%); After localization of the interlaminar space, a skin incision measuring 1.6 cm in length was performed 1.5 cm paramedially. The lumbodorsal fascia was bluntly dissected and the trocar, together with the enclosed mandrin, was gently screwed into the paraspinal muscles until the interlaminar window was reached, with the tip of the mandrin pointing medially. In this way, the paraspinal muscle attachments to the laminae and spinous processes could be well preserved in their full integrity. The mandrin was removed and the handle attached to the trocar. After a second fluoroscopy, surgery was performed with the aid of an operative microscope. After exposure of the interlaminar space, a minimal interlaminar fenestration was performed by use of drill of different size Kerrison punches, but only if necessary. Minimal partial flavectomy and bony resection. The nerve root was retracted medically and herniated disc material removed, and if necessary a partial discectomy performed.

Sasaoka (2006)24

Parallel-group RCT; Japan; High

NR

Microendoscopic discectomy N randomized: 15; N analyzed: unclear; N crossovers: 0 (0%); Intervention not described.

Microdiscectomy N randomized: 11; N analyzed: unclear; N crossovers: 0 (0%); Intervention not described.

(continued)




Author (Year)





Teli (2010)27 Parallel-group RCT; Italy; Some concerns

NR

Microendoscopic discectomy N randomized: NR; N analyzed: 70; N crossovers: 0 (0%); The Metr’X system (Medtronic Sofamor Danek, Memphis, USA) with a 16- or 18-mm tubular retractor was used. Laminotomy, medial facetectomy when needed and nerve root retraction followed by discectomy were performed.

Microdiscectomy N randomized: NR; N analyzed: 72; N crossovers: 0 (0%); Microdiscectomy with use of a surgical microscope. Laminotomy, medial facetectomy when needed and nerve root retraction followed by discectomy were performed. Discectomy N randomized: NR; N analyzed: 70; N crossovers: 0 (0%); Open discectomy with use of a magnifying loop. Laminotomy, medial facetectomy when needed and nerve root retraction followed by discectomy were performed.

(continued)




Author (Year)





Thome (2005)26 Barth (2008)99

Parallel-group RCT; Germany; Some concerns

NR

Sequestrectomy N randomized: 42; N analyzed: 42 (100%); N crossovers: 0 (0%); The spinal canal harboring the sequestrated disc material was exposed by performing a minimal interlaminar fenestration in cases of nondislocated or caudally herniated discs. In cases of cranially positioned herniated discs, a translaminar approach was undertaken, if possible. Thus, minimal removal of bone and articular structures was achieved by individualization of the procedure according to the preoperative MR imaging–depicted anatomy. In the sequestrectomy-treated group, only the herniated material was removed and the intervertebral space was not entered.

Microdiscectomy N randomized: 42; N analyzed: 42 (100%); N crossovers: 0 (0%); The spinal canal harboring the sequestrated disc material was exposed by performing a minimal interlaminar fenestration in cases of nondislocated or caudally herniated discs. In cases of cranially positioned herniated discs, a translaminar approach was undertaken, if possible. Thus, minimal removal of bone and articular structures was achieved by individualization of the procedure according to the preoperative MR imaging–depicted anatomy. In the microdiscectomy-treated group, the removal of the herniated material was followed by scalpel incision of the annulus fibrosus and resection of discal tissue from the intervertebral space—particularly the (degenerated) nucleus—with rongeurs.

Tullberg (1993)25

Parallel-group RCT; Sweden; Some concerns

NR

Microdiscectomy N randomized: 30; N analyzed: 29 (97%); N crossovers: 0 (0%); A similar dissection and disc herniation removal technique was used in both groups, the only difference being the operating microscope was used in the microdiscectomy group. Dissection involved the space between two adjacent vertebrae. The average length of the skin incision was 3.5 cm (range 2.5 to 4.5 cm). The area of exposed ligamentum flavum and laminae were the same in methods. The disc was opened by sharp dissection and all of the material was removed with rongeurs. Finally, the exposed nerve root was covered by a fat graft intended to prevent scar formation.

Discectomy N randomized: 30; N analyzed: 29 (97%); N crossovers: 0 (0%); A similar dissection and disc herniation removal technique was used in both groups, the only difference being the operating microscope was NOT used in the discectomy (standard procedure group).

(continued)




Author (Year)





Weber (1983)5

Parallel-group RCT; Norway; High

Commercial Norsk Hydro A/S

Discectomy N randomized: 60; N analyzed: 60 (100%); N crossovers: 1 (1.7%); After removal of the ligamentum flavum and in most cases a small resection of the edge of the vertebral arch above and below the exposed intervertebral space, the herniated mass of cartilage was removed extradurally. Excochleation of the disc was then performed.

Conservative management N randomized: 66; N analyzed: 66 (100%); N crossovers: 17 (25.8%); Bed rest, physiotherapy and medication for an average of 6w at a rehabilitation hospital.

Weinstein (2006)21 Weinstein (2008)45 Lurie Jon (2014)44

Parallel-group RCT; United States; High

Government; National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) and the Office of Research on Women's Health, National Institutes of Health; and by the National Institute of Occupational Safety and Health, US Centers for Disease Control and Prevention; Spine Patient Outcomes

Discectomy/microdiscectomy N randomized: 245; N analyzed N=232 in main study's primary analyses. 52w: 202 (82.47%) 2y: 186 (75.9%) 3y: 180 (73.5%) 4y: 149 (60.8%) 8y: 157 (64.1%); N crossovers Cumulative crossovers over time: 6w: 171 (69.8%) 12w: 130 (53.0%) 26w: 113 (46.1%) 52w: 107 (43.7%) 2y: 105 (42.9%) 3y: 103 (42.0%) 4y: 101 (41.2%) 8y: 97 (39.6%);

Conservative management N randomized: 256; N analyzed N=240 included in main study's primary analyses. 52w: 213 (83.2%) 2y: 187 (73.0%) 3y: 170 (66.4%) 4y: 150 (58.6%) 8y: 152 (59.4%); N crossovers Cumulative crossovers over time: 6w: 44 (17.2%) 12w: 71 (30%) 26w: 93 36.3%) 52w: 103 (40.2%) 2y: 107 (41.8% 3y: 111 (43.4%) 4y: 115 (44.9%) 8y: 122 (47.7%);

(continued)




Author (Year)





Research Trial (SPORT)

Standard open discectomy with examination of the involved nerve root. The procedure agreed on by all participating centers was performed under general or local anesthesia, with patients in the prone or knee-chest position. Surgeons were encouraged to use loupe magnification or a microscope. Using a midline incision reflecting the paraspinous muscles, the interlaminar space was entered as described by Delamarter and McCullough. In some cases the medial border of the superior facet was removed to provide a clear view of the involved nerve root. Using a small annular incision, the fragment of disk was removed as described by Spengler. The canal was inspected and the foramen probed for residual disk or bony pathology. The nerve root was decompressed, leaving it freely mobile.

The nonoperative treatment group received usual care, with the study protocol recommending that the minimum nonsurgical treatment include at least active physical therapy, education/counseling with home exercise instruction, and nonsteroidal anti-inflammatory drugs, if tolerated. Other nonoperative treatments were listed, and physicians were encouraged to individualize treatment to the patient; all nonoperative treatments were tracked prospectively.

Abbreviations: cm = centimeter; CT = computed tomography; ITT = intention-to-treat; mm = millimeter; N = number; NA = not applicable; NR = not reported; NS = not

significant; NS1 = nonsurgical intervention group; RCT = randomized controlled trial; SG1 = surgical intervention group; w = week(s); y = year(s)



Table C-2. Population characteristics of included studies

Author (Year) N eligible; N randomized Population Eligibility

Age, Mean (SD); Women, N (%); Nonwhite, N (%)

Disabled, N (%); Disability Benefits, N (%) Duration of Symptoms

Arts (2009)38 Arts (2011)43

356; 328

Inclusion: Age 18 to 70 years with sciatica due to lumbar disk herniation, which lasted more than 6 to 8w and refractory to conservative treatment; nerve root compression was confirmed with MRI. Exclusion: <1/3 of spinal canal diameter disc herniation with doubtful nerve root compression, cauda equina syndrome, previous spinal surgery at the same disc level, spondylolisthesis, central canal stenosis.

Age SG1: 41.6 (9.8) SG2: 41.3 (11.7) Female SG1: 82 (49%) SG2: 71 (45%) Nonwhite SG1: NR SG2: NR

SG1: 110 (66%) SG2: 103 (65%) Defined as: Sick leave from work; NR

Duration of symptoms, median (range) in weeks SG1: 21.0 (13 to 30) SG2: 21.0 (13 to 34)

Brouwer (2015)37 Brouwer (2017) 97

NR; 115

Inclusion: Patients between 18 and 70 years with sciatica that was refractory to conservative management for more than 6 to 8 weeks; MRI confirmation of disc herniation at the corresponding level and the herniated fragment was smaller than one-third of the spinal canal. Exclusion: Patients with cauda equina syndrome, previous spinal surgery at the same disc level, lytic or degenerative spondylolisthesis, sequestered disc herniation, disc height less than 7 mm or central canal stenosis.

Age SG1: 43.2 (11.8) SG2: 43.7 (9.7) Female SG1: 19 (35%) SG2: 24 (42%) Nonwhite NR

SG1: 26 (49%) SG2: 31 (55%) Defined as: Sick leave from work; NR

Duration of sciatica, median (range) in weeks SG1: 30.0 (9 to 182) SG2: 26.0 (8 to 260)

Chatterjee (1995)36

NR; 71 Note, the study originally planned to enroll 160 participants but was halted early because of inferiority of one treatment arm.

Inclusion: Radicular pain as dominant symptom, conventional conservative therapy for a minimum of 6 weeks, MRI-confirmed contained lumbar disc herniation at a single level, the height of which was less than 30% of the sagittal canal size. Exclusion: Dominant symptom of low back pain, MRI-confirmed disc extrusions, sequestrations, subarticular or foraminal stenosis or multiple levels of herniation.

Age SG1: 38.9 (range 20 to 56) SG2: 41.3 (range 21 to 67) Female SG1: NR (51%) SG2: NR (40%) Nonwhite NR

NR; NR

Duration of low back pain, mean (range) in weeks SG1: 78.2 (8.7 to 191.2) SG2: 143.4 (8.7 to 260.7) Duration of current episode of radicular pain, mean (range) in weeks SG1: 13 (6 to 30) SG2: 20 (6 to 38)

(continued)



Table C-2. Population characteristics of included studies (continued)





NR; 62

Inclusion: Adults with sciatica and small- to medium-sized intervertebral disk herniation (occupying less than one-third of the canal diameter at magnetic resonance [MR] imaging) that was symptomatic (leg pain with or without back pain; leg pain greater than back pain when these two coexisted; lancinating, burning, stabbing, or electrical sensation of pain; straight leg raise limited to less than 30 degrees), with the symptoms consistent with the segmental level where herniation was seen at MR imaging, conservative therapy unsuccessful. Exclusion: No neurologic deficit. Response to a 6-week course of rigorous conservative treatment; untreatable coagulopathy; active, systemic, or local infections; herniation occupying more than one-third of the spinal canal diameter and noncorrelating pain. Degenerative disease of the intervertebral disc with a disc height reduction of more than 50%–60%.

Age SG1: 38 (4.2) NS1: 36 (5.8) Female SG1: 12 (38.7%) NS1: 14 (45.2%) Nonwhite NR

NR; NR

NR

Franke (2009)34

NR; 100

Inclusion: Disc dislocation grades 3 to 5 according to Kramer et al. Exclusion: lateral disc hernia, protrusions, cauda equina syndrome, coexisting severe lumbar canal stenosis, olisthesis, scoliosis greater than 10 degrees, kyphosis greater than 15 degrees, prior lumbar spine surgeries, malignant or inflammatory disease.

Age 44 (11.7) Female 40 (40%) Nonwhite NR

NR; NR

NR

(continued)







Gerszten (2010)39

NR; 90

Inclusion: Adults between 18 and 75 years old, BMI less than 40, radicular pain score of 50 or greater as measured using a 0- to 100mm VAS, had received an epidural corticosteroid injection for the same symptoms between 3 weeks and 6 months previously with no or only partial relief and residual symptoms. Normal neurological function required and imaging evidence of a focal lumbar disc protrusion and disc height of more than 50% of that of the normal adjacent discs. In addition, the level and site of the disc protrusion had to correlate with pattern of pain. Exclusion: Extruded or sequestered disc herniation. Sciatica originating from more than one disc level, more severe axial (back) pain than radicular (leg) pain, cauda equina syndrome, progressive neurological deficit, radiological evidence of spondylolisthesis or moderate or severe stenosis at the level to be treated. History of previous spinal surgery at or adjacent to the level to be treated, spinal fracture, tumor or infection.

Age SG1: 46 (12) NS1: 42 (11) Female SG1: 24 (53%) NS1: 19 (48%) Nonwhite NR

NR; NR

Duration of symptoms, mean (range) SG1: 52w (4w to 16y) NS1: 2y (10w to 13y) P = 0.04 for SG1 vs NS1

Haines (2002)40

95; 34

Inclusion: Age 18 to 65 years with predominantly unilateral leg pain or paresthesiae, at least two of four objective signs (dermatomal sensory loss, myotomal weakness, appropriate reflex loss, appropriate nerve stretch test) and an imaging study confirming disc herniation Exclusion: No previous treatment for lumbar spinal disease, moderate or advanced lumbar spondylosis or central or lateral spinal stenosis, spondylolisthesis, progressive neurologic deficit or technical contraindications to the percutaneous procedure

Age SG1: 42.2 (12.0) SG2: 35.4 (10.1) Female SG1: 10 (47.6%) SG2: 5 (38.4%) Nonwhite SG1: 2 (9.5%) SG2: 1 (7.7%)

NR; NR

NR

(continued)







Henriksen (1996)33

99; 80

Inclusion: Age 20 to 60 years who had conservative management including bed rest, analgesics, muscle relaxers and physical therapy without sufficient improvement, diagnostic studies included positive myelograms, and/or CT scans Exclusion: Obesity, prior back surgery, or symptoms from more than one nerve root

Age Median (IQR) SG1: 39.7 (30 to 46) SG2: 42.8 (36 to 48) Female SG1: 15 (38.5%) SG2: 14 (35%) Nonwhite NR

NR; NR

NR

Hermantin (1999)41

NR; 60

Inclusion: Single intracanalicular lumbosacral (other than L1/L2) disc herniation with associated radiculopathy; a herniation not exceeding one-half of the anteroposterior diameter of the spinal canal; an absence of central or lateral osseous or ligamentous stenosis; accessibility of the disc for both arthroscopic microdiscectomy and laminotomy; failure to respond to nonoperative measures for 14 weeks; more pain in the lower extremities than in the back; the presence of positive tension signs with or without an accompanying neurological deficit; a dermatomal distribution of pain in the lower extremities matching that seen on imaging studies and specific nerve-root involvement. Exclusion: previous operation on the low back, litigation or Workers' Compensation claim involving the disc herniation, central or lateral stenosis of the spinal canal, severe degenerative narrowing of the intervertebral disc space at the index level, evidence on imaging of global bulging of the intervertebral disc associated with central or lateral stenosis; a sequestered herniation that had migrated, a large central or extraligamentous herniation between L5/S1.

Age SG1: Mean 40 (range 18-67) SG2: Mean 39 (range 15-66) Female SG1: 13 (43.4%) SG2: 8 (26.7%) Nonwhite NR

NR; NR (this is presumably 0 (0%) as patients with any litigation or Workers' Compensation claim involving the disc herniation are excluded)

Minimum duration of nonoperative treatment prior to randomization was 14w in both groups.

(continued)







Huang (2005)23 NR; 21 Note: 22 patients were enrolled, but only 21 underwent randomization; one patient insisted on assignment to SG2.

Inclusion: Age criteria NR, symptomatic herniated intervertebral discs who were scheduled to undergo elective lumbar discectomy; 16 patients failed to respond to conservative treatment after three months, and six patients with acute attack of intractable back and leg pain that demonstrated no improvement after 1-2 weeks of absolute bed rest. Exclusion: Recurrent lumbar disc herniation, significant motor deficit or sphincter disturbance

Age SG1: 39.2 (10.8) SG2: 39.8 (11.0) Female SG1: 4 (40%) SG2: 3 (25%) Nonwhite NR

NR; NR

NR

Malter (1996)42 2,175; NA

Patients younger than 65 years old identified from the MEDSTAT commercially-available database of non-governmental insurers from all 50 states. Inclusion: Diagnosis of a herniated intervertebral disc confirmed with imaging. Exclusion: NR

Age SG1: 46 (NR) NS1: 46 (NR) Female SG1: 208 (56) NS1: 829 (46) Nonwhite NR

NR; NR

NR

(continued)







Mayer (1993)32 NR; 40

Inclusion: Clinical symptoms due to discogenic lumbar nerve root compression with radicular symptoms such as positive straight-leg raising test, sciatica, sensory disturbances, mild motor weakness, and/or reflex differences, failed conservative therapy, "contained" (when the outer border of the annulus fibrosus was still intact.) or small non-contained ( extrusion of nucleus pulposus under the posterior longitudinal ligament but still at the level of the disc space and occupying not more than one-third of the sagittal diameter of the spinal canal), imaging confirmation. Exclusion: Severe motor deficits, conus or cauda equina syndrome, or rapidly progressing neurological symptoms, patients with signs of segmental instability or previous surgery at the same site, worker's compensation claims, malformations, tumors, or posttraumatic root compression, large "non-contained" disc herniations extending cranially or caudally to the level of the disc space, spinal stenosis, or spondylolisthesis.

Age SG1: 39.8 (10.4) SG2: 42.7 (10) Female SG1: 8 (40%) SG2: 6 (30%) Nonwhite NR

NR; NR (this is presumably 0 (0%) as patients with worker's compensation claims are excluded)

Duration of symptoms, mean (SD) in weeks SG1: 27.6 (NR) SG2: 29.2 (NR)

(continued)







McMorland (2010)22

60; 40

Inclusion: Unilateral radiculopathy secondary to lumbar disc herniation at L3/4, L4/5, or L5/S1, with leg-dominant symptoms with objective signs of nerve root tethering ± neurologic deficit correlated with evidence of appropriate root compression on magnetic resonance imaging, failed at least 3 months of nonoperative management including treatment with analgesics, lifestyle modification, physiotherapy, massage therapy, and/or acupuncture. Exclusion: major neurological deficits (Cauda equina syndrome, rapidly progressing neurological symptoms (e.g. foot drop)), previous surgery at symptomatic level, concurrent treatment involving spinal manipulation at time of enrollment, prolonged use of systemic corticosteroids, osteopenia/osteoporosis, spondylolisthesis grade III or IV.

Age SG1 male: 42.85 (NR) SG1 female: 40.1 (NR) NS1 male: 36.4 (NR) NS1 female: 48.33 (NR) Female SG1: 7 (35%) NS1: 9 (45%) Nonwhite NR

SG1: 9 (45%) NS1: 11 (55%) Defined as: "Medical leave" for work status; SG1: 1 (5%) NS1: 1 (5%) N (%) receiving 3rd party disability insurance SG1: 0 (0%) NS1: 0 (0%)

N (%) with duration of complaint 12-26w SG1: 3 (15%) NS1: 6 (30%) N (%) with duration of complaint 26-52w SG1: 5 (25%) NS1: 6 (30%) N (%) with duration of complaint >52w SG1: 12 (60%) NS1: 8 (40%)

(continued)







North (2005)47 99; 60 (randomized) 50 (treated) Of the 10 randomized but not treated, 9 did not receive authorization from Workers Compensation for study participation and 1 had a stroke prior to treatment). Patients with Worker's Compensation consented to randomization as often as other patients.

Inclusion: Surgically remediable nerve root compression and concordant complaints of persistent or recurrent radicular pain, with or without low back pain, after one or more lumbosacral spine surgeries. Pain refractory to conservative care, with concordant neurological, tension, and/or mechanical signs and imaging findings of neural compression. Exclusion: A disabling neurological deficit (e.g., foot drop, neurogenic bladder) in the distribution of a nerve root or roots caused by surgically remediable compression; radiographically demonstrated (by myelographic block or its magnetic resonance imaging equivalent) critical cauda equina compression; radiographic evidence of gross instability (spondylolisthesis or abnormal subluxation); unresolved issues of secondary gain; a chief complaint of axial (low back) pain exceeding radicular (hip, buttock, and leg) pain.

Age 52.0 (13.5) Female 26 (52%) Nonwhite NR

NR; Overall 15 (30%) receiving workers compensation

Mean (SD) number of prior operations SG1: 2.5 (1.1) NS1: 2.5 (1.1)

(continued)







Osterman (2003)31

NR; 56

Inclusion: Age 20 to 50 years with 1) below knee radicular pain of 6 to 12 weeks’ duration, 2) a CT finding of intervertebral disc extrusion or sequester, 3) at least one specific physical finding (a positive straight leg raising test, muscle weakness, altered reflexes, dermatomal sensory change). Exclusion: 1) previous back surgery, 2) spondylolisthesis, 3) symptomatic spinal stenosis, 4) over 3 months’ continuous sick leave because of low back pain or leg pain, 5) a condition confounding evaluation of treatment outcomes (vascular claudication, symptomatic osteoarthritis, previous major trauma, diabetic polyneuropathy), or 6) a contraindication to conservative treatment (cauda equina syndrome, progressive neurologic deficit, or intolerable pain).

Age SG1: 37 (7) NS1: 38 (7) Female SG1: 13 (46.4%) NS1: 9 (32.1%) Nonwhite NR

NR; NR

Duration of leg pain, mean (SD) in weeks SG1: 11.0 (4.6) NS1: 8.6 (3.0) Duration of back pain, mean (SD) in weeks SG1: 13.4 (6.7) NS1: 10.4 (4.6)

Peul (2007)30 Peul (2008)95 Lequin (2013)96

599; 283

Inclusion: Eligible patients were 18 to 65 years of age, had a radiologically confirmed disk herniation, and had received a diagnosis from an attending neurologist of an incapacitating lumbosacral radicular syndrome that had lasted for 6 to 12 weeks with correlation of MRI findings with symptoms. Exclusion: Cauda equina syndrome, muscle paralysis, insufficient strength to move against gravity, occurrence of another episode of symptoms similar to those of the current episode during the previous 12 months, previous spine surgery, bony stenosis, and spondylolisthesis.

Age SG1: 41.7 (9.9) NS1: 43.4 (9.6) Female SG1: 52 (37%) NS1: 45 (32%) Nonwhite NR

SG1: 107 (76%) NS1: 116 (82%) Defined as: Sick leave from work; NR

Duration of symptoms, mean (SD) in weeks SG1: 9.43 (2.37) NS1: 9.48 (2.11)

(continued)







Ruetten (2008)29

NR; 200

Inclusion: Clinically symptomatic disc herniation with radicular pain and neurologic deficits. Exclusion: NR

Age 43 (range 20 to 68) Female 116 (58%) Nonwhite NR

NR; NR

Duration of symptoms, mean (range) in weeks 11.71 (0.14 to 68) 162 of the 200 patients had received a mean of 9w of conservative treatment.

Ruetten (2009)48

NR; 100

Inclusion: Previous conventional discectomy with acute occurrence of radicular leg symptoms on the same side after a pain-free interval and who showed a recurrent disc herniation in the same level with MRI. Inclusion criteria specific for the full endoscopic transforaminal access were (1) sequestering of material located cranially below the lower edge of the cranial pedicle or caudally not over the middle of the caudal pedicle and (2) lateral radiologic evidence that the foramen was not overlaid by the pelvis beyond the middle of the cranial pedicle. Exclusion: NR

Age 39 (range 23 to 59) Female 44 (44%) Nonwhite NR

NR; NR

Duration of symptoms, mean (range) in weeks 9.85 (0.14 to 56)

(continued)







Ryang (2008)28 Gempt (2013)98

NR; 60

Inclusion: 1) single level virgin lumbar disc herniation; 2) typical monoradicular symptoms attributable to the involved lumbar segment with predominant sciatica compared to less severe lower back pain; and 3) failure of 8 to 12 w of conservative treatment, intolerable sciatica, or rapidly progressive neurological deficits including motor deficits, bladder dysfunction, and cauda equina syndrome. Exclusion: (1) history of previous lumbar back surgery or conservatively treated lumbar disc herniation at adjacent levels; (2) signs of spinal instability or other spinal abnormalities such as bone disease, spinal infection, malignancy, or signs of spinal canal stenosis on computed tomography or magnetic resonance imaging and neurogenic claudication; (3) intra- and extraforaminal far lateral disc herniation; (4) chronic pain syndrome and opioid abuse; (5) pending worker's compensation.

Age SG1: 38.2 (9.3) SG2: 39.1 (11.3) Female SG1: 17 (56.7%) SG2: 11 (36.7%) Nonwhite NR

NR; NR

NR

Sasaoka (2006)24

NR; 33

NR Age 42.4 (range 20 to 72) SG1: 36.5 (range 25 to 60) SG2: 37.7 (range 20 to 58) Female 14 (42.4%) SG1: 9 (60.0%) SG2: 3 (27.3%) Nonwhite NR

NR; NR

NR

(continued)







Teli (2010)27 NR; 240

Inclusion: Aged 18 to 65 years, symptomatic, single level posterior lumbar disc herniation with diagnosis made by spine specialists (orthopaedic and neurosurgeons) with pain and/or neurological signs in concordant distribution lasting at least 6 weeks despite appropriate conservative treatment consisting of systemic drugs for pain relief and/or epidural steroid administration, imaging confirmation with MRI or a CT scan of the lumbar spine, supplemented with plain X-rays of the lumbar spine including the thoracolumbar tract to exclude or confirm the presence of a segmentation anomaly. Exclusion: cauda equina symptoms, foraminal or extra-foraminal herniations, cervical or lumbar spine stenosis of any etiology, malignancy, previous spine surgery, spinal deformity including spondylolisthesis, concurrent infection and rheumatic disease.

Age 39.3 (range 27 to 61) Female 73 (34.4% of N analyzed) Nonwhite NR

NR; NR

Duration of pain, mean (SD) in weeks SG1: 11 (5) SG2: 12 (6) SG3: 11 (5)

Thome (2005)26 Barth (2008)99

221; 84

Inclusion: Age between 18 and 60 years, with MRI- documented intraspinal (not extraforaminal) disc fragment that had perforated the annulus fibrosus. Exclusion: Previous lumbar spine surgery, emergency indication for surgery, MRI-documented lumbar spinal stenosis or spondylolisthesis.

Age SG1:42 (9) SG2: 40 (10) Female SG1: 18 (42.9%) SG2: 19 (45.2%) Nonwhite NR

NR; NR

Duration of symptoms, mean (SD) in weeks SG1: 11 (12) SG2: 8 (10) P=0.27 for SG1 vs SG2

(continued)







Tullberg (1993)25

NR; 60

Inclusion: Sciatica, failed 2 months of conservative therapy and had CT-confirmed diagnosis of single lumbar disc herniation within the spinal canal Exclusion: Recurrent disc herniation or previous back surgery

Age SG1: 40 (range 17 to 59) SG2: 38 (range 18 to 64) Female SG1: 12 (40%) SG2: 9 (30%) Nonwhite NR

N (%) with specified duration of disablement: SG1 none: 6 (NR) <4w: 5 (NR) 4w to 12w: 8 (NR) 16w to 26w: 5 (NR) 27w to 52w: 5(NR) >1y: 1 (NR) SG2 <4w: 0 (NR) 4w to 12w: 6 (NR) 16w to 26w: 7 (NR) 27w to 52w: 13 (NR) >1y: 4 (NR) Defined as: Sick leave from work; NR

N (%) with specified duration of symptoms: SG1 <4w:1 (NR) 4w to 12w: 7 (NR) 16w to 26 w: 2 (NR) 27w to 52w: 14 (NR) >52w: 6 (NR) SG2 <4w: 0 (NR) 4w to 12w: 6 (NR) 16w to 26w: 7 (NR) 27w to 52w: 13 (NR) >1y: 4 (NR)

Weber (1983)5 NR; 126

Inclusion: Patients admitted to hospital with sciatica, with continued radicular pain provoked by moderate exercise, sitting position, or increased abdominal pressure after an initial 14d period of observation, radiographic confirmation of definite or possible disc herniation based on radiculography with water soluble contrast. Exclusion: Definite indications for surgery (intolerable pain, suddenly occurring or progressive muscle weakness, bladder or rectum paresis), severe or immobile scoliosis; patients with moderate symptoms but who showed signs of continuous improvement as a result of bed rest, physiotherapy, or medication during the 14d observation period were also excluded.

Age SG1: 40.0 (NR) NS1: 41.7 (NR) Female SG1: 28 (46.7%) NS1: 30 (45.5%) Nonwhite NR

NR; NR

NR

(continued)







Weinstein (2006) 21 [SPORT] Weinstein (2008)45 Lurie Jon (2014)44

1991; 501

Inclusion:18 years and older and diagnosed as having intervertebral disk herniation (confirmed by MRI or CT) and persistent symptoms (radicular pain and evidence of nerve-root irritation with a positive nerve-root tension sign or corresponding neurologic deficit) despite nonoperative treatment for >6 weeks. Exclusion: Prior lumbar surgery, cauda equina syndromes, scoliosis greater than 15 degrees, segmental instability, vertebral fractures, spine infection of tumor, inflammatory spondyloarthropathy.

Age Overall: 42.3 (11.6) SG1: 41.7 (11.8) NS1: 43.0 (11.3) Female Overall: 194 (41.1%) SG1: 101 (44%) NS1: 93 (39%) Nonwhite Overall: 73 (15.5%) SG1: 35 (15.1%) NS1: 38 (15.8%)

Overall: 58 (12.3%) SG1: 27 (12%) NS1: 31 (13%) Defined as: Employment status is "Disabled"; Overall: 76 (16.1%) SG1: 36 (16%) NS1: 40 (17%)

NR

Abbreviations: CI = confidence interval; CT = computed tomography; d = days; mm = millimeter; MRI = magnetic resonance imaging; N = number; NR = not reported; NS = not

significant; NS1 = nonsurgical intervention group; SD = standard deviation; SE = standard error; SG1 = surgical intervention group; VAS = visual analogue scale; w = week(s); y

= year(s)



Table C-3. Individual study findings related to efficacy outcomes Part 1 (pain, neurological symptoms, quality of life)

Main Study Author (Year) Follow-up Studies Author (Year); Risk of Bias

Intervention Names N randomized N analyzed Pain Neurologic Symptoms Quality of Life

Arts (2009)38 Arts (2011)43 Low

SG1: Tubular discectomy N randomized: 167 N analyzed: 166 (99.4%) for primary analyses SG2: Microdiscectomy N randomized: 161 N analyzed: 159 (98.8%) for primary analyses

VAS 100mm leg pain, mean (SD) Baseline: SG1 62.6 (21.1); SG2 61.7 (24.0) AMD (95% CI); [Negative AMD favors SG1] 4w: 4.5 (−0.3 to 9.3) 8w: 4.5 (−0.4 to 9.3) 26w:2.0 (−2.9 to 6.8) 52w: 4.4 (−0.5 to 9.4) 2y: 1.3 (-3.6 to 6.2) RM 1w to 52w: 4.2 (0.9 to 7.5), P=0.01 main treatment effect, P=0.12 treatment X time interaction RM 1w to 2y: 3.3 (0.2 to 6.2), P=0.04 main treatment effect, P=0.08 treatment X time interaction VAS 100mm back pain, mean (SD) Baseline: SG1 40.2 (27.0); SG2 38.3 (27.8) AMD (95% CI); [Negative AMD favors SG1] 4w: 3.1 (−1.9 to 8.1) 8w: 3.8 (−1.3 to 8.8) 26w: 3.5 (−1.5 to 8.6) 52w: 4.9 (−0.2 to 10.1) 2y: 4.1 (-1.2 to 9.4) RM 1w to 52w: 3.5 (0.1 to 6.9), P=0.04 main treatment effect, P=0.37 treatment X time interaction RM 1w to 2y: 3.0 (-0.2 to 6.3), P=0.07 main treatment effect, P=0.05 treatment X time interaction SF-36 bodily pain, mean (SD) Baseline: SG1 27.8 (18.2); SG2 25.2 (17.7) AMD (95% CI); [Positive AMD favors SG1] 4w: -1.6 (−6.7 to 3.6) 8w: -5.1 (−10.3 to 0.1) 26w: -4.9 (−10.0 to 0.3) 52w: -3.8 (−9.0 to 1.5) 2y: -3.2 (-8.6 to 2.3)

NR

NR

(continued)



Table C-3. Individual study findings related to efficacy outcomes Part 1 (pain, neurological symptoms, quality of life) (continued)



Arts (2009)38 Arts (2011)43 Low (continued)

RM 1w to 52w: -3.3 (-7.3 to 0.7), P=0.10 main treatment effect, P=0.28 treatment X time interaction RM 1w to 2y: -2.8 (-6.7 to 1.0), P=0.14 main treatment effect, P=0.22 treatment X time interaction Sciatica index, Bothersomeness Score, mean (SD) Baseline: SG1 14.1 (4.8); 14.2 (5.0) AMD (95% CI); [Negative AMD favors SG1] 4w: 0.3 (−0.7 to 1.4) 8w: 0.8 (−0.3 to 1.8) 26w: 1.1 (0 to 2.1) 52w: 0.9 (−0.1 to 2.0) 2y: 0.2 (-0.8 to 1.3) RM 1w to 52w: 0.7 (-0.1 to 1.5), P=0.10 main treatment effect, P=0.37 treatment X time interaction RM 1w to 2y: 0.5 (-0.3 to 1.3), P=0.26 main treatment effect, P=0.40 treatment X time interaction Sciatica index, Frequency Score, mean (SD) Baseline: SG1 16.0 (4.4); SG2 15.5 (4.3) AMD (95% CI); [Negative AMD favors SG1] 4w: 0.3 (-0.8 to 1.4) 8w: 0.8 (-0.4 to 1.9) 26w: 1.0 (-0.1 to 2.1) 52w: 1.0 (-0.1 to 2.2) 2y: 0.3 (-0.9 to 1.5) RM 1w to 52w: 0.7 (-0.2 to 1.7), P=0.14 main treatment effect, P=0.41 treatment X time interaction RM 1w to 2y: 0.5 (-0.5 to 1.4, P=0.32 main treatment effect, P=0.45 treatment X time interaction

(continued)






Brouwer (2015)37 Brouwer (2017)97 Some concerns

SG1: Percutaneous laser disc decompression N randomized: 57 N analyzed: 55 (96.5%) SG2: Microdiscectomy, N randomized: 58 N analyzed: 57 (98.3%)

VAS 100mm leg pain, mean (SD) Baseline: SG1 56.9 (20.4); SG2 60.7 (19.9) AMD (95% CI); [Negative AMD favors SG1] 4w: 7.4 (-1.9 to 16.8) 8w: 5.7 (-3.7 to 15.0) 26w: 4.2 (-5.2 to 13.6) 52w: 5.7 (-3.8 to 15.2) 2y: -2.7 (-12.3 to 6.8) RM 1w to 52w: 6.9 (1.3 to 12.6) RM 1w to 2y: 5.0 (-0.2 to 10.2), P=0.06 main treatment effect, P=0.42 treatment X time interaction VAS 100mm back pain, mean (SD) Baseline: SG1 44.7 (27.6); SG2 45.8 (26.7) AMD (95% CI); [Negative AMD favors SG1] 4w: -2.0 (-11.3 to 7.2) 8w: 6.3 (-2.9 to 15.5) 26w: 9.4 (0.1 to 18.6) 52w: 7.6 (-1.7 to 16.9) 2y: -1.5 (-11.0 to 8.0) RM 1w to 52w: 4.6 (-1.1 to 10.4) RM 1w to 2y: 3.0 (-2.2 to 8.1), P=0.26 main treatment effect, P=0.58 treatment x time interaction SF-36 bodily pain, mean (SD) Baseline: SG 1 32.8 (20.5); SG2 30.0 (16.1) AMD (95% CI); [Positive AMD favors SG1] 4w: 4.1 (-4.8 to 12.9) 8w: 0.6 (-8.1 to 9.3) 26w: -11.3 (-20.1 to -2.4) 52w: -2.5 (-11.3 to 6.4) 2y: 2.2 (-7.0 to 11.4) RM 1w to 52w: -1.6 (-7.3 to 4.2)

NR

NR

(continued)






Brouwer (2015)37 Brouwer (2017)97 Some concerns (continued)

RM 1w to 2y: -0.3 (-5.6 to 5.0), P=0.91 main treatment effect, P=0.14 treatment x time interaction Note: the scoring of the following measures was adapted by the study authors, instead of summing items to report a score range of 0 to 24, the study authors reported the average per item, which can range from 1 to 4. Sciatica index, Bothersomeness, mean (SD) Baseline: SG1 3.3 (1.2); SG2 3.1 (1.3) AMD (95% CI); [Negative AMD favors SG1] 4w: 0.1 (-0.4 to 0.6) 8w: 0.2 (-0.2 to 0.7) 26w: 0.3 (-0.2 to 0.7) 52w: 0.2 (-0.2 to 0.7) 2y: -0.1 (-0.6 to 0.4) RM 1w to 52w: 0.2 (-0.1 to 0.5) RM 1w to 2y: 0.1 (-0.2 to 0.4), P=0.56 main treatment effect, P=0.89 treatment x time interaction Sciatica index, Frequency, mean (SD) Baseline: SG1 3.6 (1.1); SG2 3.8 (1.2) AMD (95% CI); [Negative AMD favors SG1] 4w: 0.1 (-0.4 to 0.6) 8w: 0.1 (-0.4 to 0.5) 26w: 0.1 (-0.4 to 0.6) 52w: 0.1 (-0.3 to 0.6) 2y: 0 (-0.4 to 0.3) RM 1w to 52w: 0.1 (-0.2 to 0.4) RM 1w to 2y: -0.2 (-0.7 to 0.3), P=0.92 main treatment effect, P=0.92 treatment x time interaction

(continued)






Chatterjee (1995)36 Some concerns

SG1: Automated Percutaneous Lumbar Discectomy (APLD) N randomized: 31 N analyzed: 31 (100%) SG2: Microdiscectomy N randomized: 40 N analyzed: 40 (100%)

NR NR NR

Erginousakis (2011)35 High

SG1: Percutaneous Disc Decompression N randomized: 31 N analyzed: 31 (100%) NS1: Conservative management N randomized: 31 N analyzed: 31 (31%)

VAS 10 cm pain, mean (SD) Baseline: SG1 7.4 (1.4); NS1: 6.9 (1.9), P NR 12w: SG1 3.0 (2.4); NS1 0.9 (2.0); P>0.005 (described in text as NS) 52w: SG1 1.7 (2.4); NS1 4.0 (3.4); P=0.005 2y: SG1 1.6 (2.5); NS1 4.1 (3.4); P=0.004 Mean % pain reduction at 2y SG1: 86% NS1: 36% N (%) with category of pain reduction at 2y 100% pain relief: SG1 17 (55%); NS1 6 (19%); Calculated P=0.008 50% pain relief: 4 (13%); NS1 2 (6%) 0% pain relief: 2 (6%); NS1 3 (10%) Aggravation of pain: SG1 0 (0%); NS1 2 (6%)

NR NR

(continued)






Franke (2009)34 Some concerns

SG1: Microscopically assisted percutaneous nucleotomy (MAPN) N randomized: 52 N analyzed: 52 (100%) SG2: Microdiscectomy N randomized: 48 N analyzed: 48 (100%)

VAS (sum of leg and back) pain (RM at 52w) Significant within group reduction in pain over time in both groups, P < 0.001 Conflicting between group differences depending on which of the two centers the procedure was performed, P=0.006 at one center, P=0.7 at other center Post hoc analysis at one of the two centers VAS back pain 8w: Larger decrease in SG1, P=0.002 26w: Larger decrease in SG1, P=0.003 52w: No difference, P=0.467 No difference was found for the VAS leg pain at any time points No difference in VAS back or leg pain at the other center at any time point.

Overall at 52w (NR by group): 83% of motor deficits resolved completely 68% of sensory deficits resolved completely

NR

(continued)






Gerszten (2010)39 Some concerns (6w outcomes) High (12w and later outcomes)

SG1: Plasma disc decompression using coblation technology (PDD) N randomized: 46 N analyzed: 29 (64% of ITT sample) at 26w NS1: Transforaminal epidural steroid injection (TFESI) N randomized: 44 N analyzed: 28 (70% of ITT sample) at 26w

VAS 100 mm leg pain, mean (SD) Baseline: SG1: 72 (13); NS1 75 (14) Change in score, mean (SE); [larger negative change favors SG1]

6w: SG1 -42 (5); NS1 -21 (4); P=0.002 Calculated AMD -21 12w: SG1 -46 (4); NS1 -23 (5); P=0.0001 Calculated AMD -23 26w: SG1 -47 (6); NS1 -21 (5); P=0.0008 Calculated AMD -26 VAS 100 mm back pain, mean (SD) Baseline: SG1 44 (24); NS1 53 (23) Change in score, mean (SE); [larger negative change favors SG1] 6w: SG1 -18 (4); NS1 1 (3); P=0.0005 Calculated AMD -19 12w: SG1 -17 (5); NS1 7 (4); P=0.0001 Calculated AMD -24 26w: SG1 -21 (5); NS1 -0.4 (4); P=0.002 Calculated AMD -21 SF-36 bodily pain at 26w Larger improvement in SG1 compared with NS1, P=0.0039

N (%) with full muscle strength at 6w No significant difference between groups on left or right side at L3, L4, L5 or S1 (8 comparisons) N (%) with normal tactile sensitivity at 6w Right side S1 (p=0.01) SG1: 39 (98%) NS1: 25 (78%) All other 7 comparisons NS.

NR

(continued)






Haines (2002)40 High

SG1: Automated percutaneous discectomy, endoscopic percutaneous discectomy (APD/EPD) N randomized: 21 N analyzed: 17 (81.0%) at 26w SG2: Discectomy N randomized: 13 N analyzed: 10 (76.9%) at 26w

NR NR NR

Henriksen (1996)33 Some concerns

SG1: Microdiscectomy N randomized: 40 N analyzed: 39 (97.5%) SG2: Standard discectomy N randomized: 40 N analyzed: 40 (100%)

VAS 100 mm leg pain at 4w and 6w Actual values only depicted in a figure and variance NR, no differences between the groups reported VAS 100 mm back pain at 4w and 6w Actual values only depicted in a figure and variance NR, no differences between the groups reported

NR NR

(continued)






Hermantin (1999)41 Some concerns

SG1: Video-assisted arthroscopic microdiscectomy N randomized: 30 N analyzed: 30 (100%) SG2: Discectomy, with laminotomy N randomized: 30 N analyzed: 30 (100%)

VAS 10 cm pain, mean (SD) Baseline: SG1 6.8 (NR); SG2 6.6 (NR) Unspecified follow-up time: SG1: 1.2 (NR); SG2: 1.9 (NR)

Follow-up time for these measures were NR N (%) with postoperative reflex abnormalities SG1: 7 (20.3%) SG2: 6 (20.0%) Calculated P=1.0 N (%) with sensory deficits SG1: 16 (53.3%) SG2: 18 (60.0%) Calculated P=0.79 N (%) with motor weakness SG1: 5 (16.7%) SG2: 10 (33.3%) Calculated P=0.23

NR

Huang (2005)23 Some concerns

SG1: Microendoscopic discectomy N randomized: 10 N analyzed: 10 (100%) SG2: Discectomy N randomized: 12 N analyzed: 12 (100%)

NR NR NR

(continued)






Mayer (1993)32 High

SG1: Percutaneous endoscopic discectomy N randomized: 20 N analyzed: 20 (100%) SG2: Microdiscectomy N randomized: 20 N analyzed: 20 (100%)

N (%) of patients with low back pain Baseline: SG1 19 (95%); SG2 20 (100%) 2y: SG1 9 (47.4%); SG2 4 (20%) Calculated P=0.18 N (%) of patients with sciatica Baseline: SG1 20 (100%); SG2 20 (100%) 2y: SG1 4 (20.0%); SG2 7 (35%) Calculated P=0.48

N (%) of patients with sensory deficit Baseline: SG1 13 (65%); SG2 16 (80%) 2y: SG1 1 (5%); SG2 5 (25%); Calculated P=0.18 N (%) of patients with motor deficit Baseline: SG1 1 (5%); SG2 4 (20%) 2y: SG1 0 (0%); SG2 0 (0%); Calculated P=1.0 N (%) of patients with reflex differences Baseline: SG1 10 (50%); SG2 7 (35%) 2y: SG1 2 (10%); SG2 2 (10%); Calculated P=1.0

NR

McMorland (2010)22 Some concerns

SG1: Microdiscectomy N randomized: 20 N analyzed: 12w 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w 20 (100%) 52w 15 (75%) NS1: Spinal manipulation N randomized: 20 N analyzed: 12w 20 (100%) (outcomes reported only to 12w for ITT analysis) 24w 20 (100%) 52w 17 (85%)

SF-36 bodily pain, mean (SD) Baseline: SG1 27.3 (19.8); NS1 28.5 (21.8) 6w: SG1 41.4 (24.1); NS1 45.8 (21.3) 12w: SG1 57.4 (22.3); NS1 47.1 (18.4) RM 0w to 12w: AMD NR, P=0.031 for time effect, P=0.341 for main treatment effect, P=0.367 for treatment X time interaction McGill Pain Questionnaire, Pain Rating Intensity, mean (SD) Baseline: SG1 32.5 (12.9); NS1 28.7 (17.4) 6w: SG1 18.4 (16.3); NS1 21.7 (13.7) 12w: SG1 13.0 (16.3); NS1 19.4 (14.3) RM 0w to 12w: AMD NR, P=0.013 for time effect, P=0.103 for main treatment effect, P=0.754 for treatment X time interaction

NR SF-36 Total Score, mean (SD) Baseline: SG1 379.5 (149.8); NS1: 381.3 (161.9) 6w: SG1 429.1 (157.3); NS1 445.6 (142.8) 12w: SG1 500.3 (179.7); NS1: 484.6 (148.9) 0 to 12w (RM), AMD (95%CI) [positive AMD favors SG1] AMD NR, P=0.016 for time effect, P=0.382 for main treatment effect, P=0.683 for treatment X time interaction

(continued)






McMorland (2010)22 Some concerns (continued)

McGill Pain Questionnaire, Number of words chosen, mean (SD) Baseline: SG1 13.2 (5.0); NS1 12.0 (5.5) 6w: SG1 8.8 (6.4); NS1 10.8 (6.1) 12w: SG1 5.7 (5.1); NS1 9.6 (6.3) RM 0w to 12w: AMD NR, P=0.029 for time effect, P=0.080 for main treatment effect, P=0.574 for treatment X time interaction McGill Pain Questionnaire, Present pain intensity, mean (SD) Baseline: SG1 2.7 (1.0); NS1 2.4 (0.8) 6w: SG1 1.6 (1.3); NS1 1.8 (0.7) 12w: SG1 1.5 (1.3); NS1 1.6 (0.9) RM 0w to 12w: AMD NR, P=0.010 for time effect, P=0.094 for main treatment effect, P=0.736 for treatment X time interaction Aberdeen back pain scale, mean (SD) Baseline: SG1 45.1 (17.8); NS1 44.7 (12.9) 6w: SG1 32.3 (22.2); NS1 34.8 (18.6) 12w: SG1 25.8 (23.7); NS1 35.6 (18.9) RM 0w to 12w: AMD NR, P= 0.017 for time effect, P=0.034 for main treatment effect (favors SG1), P=0.836 for treatment X time interaction

(continued)






North (2005)47 High

SG1: Repeat lumbosacral decompression N randomized: 26 N analyzed: 26 (100%) NS1: Spinal cord stimulation N randomized: 24 N analyzed: 19 (79.2%)

NR NR NR

(continued)






Osterman (2003)31 High

SG1: Microdiscectomy N randomized: 28 N analyzed: 6w 26 (93%) 12w 26 (93%) 26w 26 (93%) 52w 21 (75%) 2y 26 (93%) NS1: Physiotherapy N randomized: 28 N analyzed: 6w 26 (93%) 12w 26 (93%) 26w 22 (78.6%) 52w 20 (71.4%) 2y 24 (86%)

VAS 100 mm leg pain, mean (SD) Baseline: SG1 61 (20); NS1 57 (21) [Negative AMD favors SG1] 6w: SG1 12 (20); NS1 25 (27) Calculated AMD -17 (adj. for baseline) 12w: SG1 9 (16); NS1 16 (25) Calculated AMD -11 (adj. for baseline) 26w: SG1 9 (20); NS1 18 (29) Calculated AMD -13 (adj. for baseline) 52w: SG1 6 (11); NS1 9 (19) Calculated AMD -7 (adj. for baseline) 2y: SG1 6 (11); NS1 15 (24) Calculated AMD -13 (adj. for baseline) RM 0 to 2y: AMD -9 (95% CI, -20 to 1) VAS 100 mm back pain, mean (SD) Baseline: SG1 53 (25); NS1 47 (28) [Negative AMD favors SG1] 6w: SG1 21 (25); NS1 28 (24) Calculated AMD -13 (adj. for baseline) 12w: SG1 15 (20); NS1 22 (23) Calculated AMD -13 (adj. for baseline) 26w: SG1 13 (22); NS1 20 (28) Calculated AMD -13 (adj. for baseline) 52w: SG1 19 (25); NS1 17 (23) Calculated AMD -4 (adj. for baseline) 2y: SG1 11 (18); NS1 21 (27) Calculated AMD -16 (adj. for baseline) RM 0 to 2y: AMD (95% CI) -7 (-17 to 3)

N (%) with muscle weakness 6w: SG1 14 (53.8%); NS1 12 (46.2%) 12w: SG1 11 (42.3%); NS1 12 (46.2%) 26w: NR 52w: SG1 6 (28.6%); NS1 6 (30%) 2y: NR

15D Health-related quality of life, mean (SD) Baseline: SG1 0.83 (0.07); NS1 0.84 (0.06) 6w: SG1 0.92 (0.07); NS1 0.89 (0.09) 12w: SG1 0.94 (0.06); NS1 0.91 (0.09) 26w: SG1 0.95 (0.06): NS1 0.90 (0.13) 52w: SG1 0.95 (0.05); NS1 0.94 (0.07) 2y: SG1 0.95 (0.08); NS1 0.93 (0.12) RM 0 to 2y, AMD (95% CI) [positive AMD favors SG1] -0.03 (-0.07 to 0.01)

(continued)






Peul (2007)30 Peul (2008)95 Lequin (2013)96 High

SG1: Microdiscectomy N randomized: 141 N analyzed: 1y: 140 (99.3%) 2y: 130 (92.2%) 5y: 115 (81.6%) NS1: Conservative management N randomized: 142 N analyzed: 1y: 141 (99.3%) 2y: 130 (91.5%) 5y: 116 (81.7%)

VAS 100 mm leg pain, mean (SD) Baseline: SG1 67.2 (27.7); NS1 64.4 (21.2) AMD (95% CI); [negative AMD favors SG1] 8w: -17.7 (-23.1 to -12.3) 26w: -6.1 (-10.0 to -2.2) 52w: 0 (-4.0 to 4.0) 2y: 2 (-2.0 to 6.0) 5y: 2.7 (-2.9 to 8.4) Cumulative score on VAS 100 for leg pain 0 to 52w SG1: AUC 635.3 (SE 58.6) NS1: AUC 977.0 (SE 68.3) AMD: -341.7 (95% CI, -519.6 to 163.8) Cumulative score on VAS 100 for leg pain 0 to 2 years SG1: AUC 1,110.2 (SE 133.3) NS1: AUC 1,487.1 (SE 137.7) AMD: -376.8 (95% CI, -754.6 to 0.9) Cumulative score on VAS 100 for leg pain 0 to 5 years SG1: NR NS1: NR Reported as no significant difference VAS 100 mm back pain, mean (SD) Baseline: SG1 33.8 (29.6); NS1 30.8 (27.7) AMD (95% CI); [negative AMD favors SG1] 8w: -11.3 (-17.4 to -5.6) 26w: -2.3 (-8.2 to 3.6) 52w: -2.3 (-8.2 to 3.6) 2y: -1.4 (-6.3 to 4.5) 5y: 3.1 (-4.2 to 10.3)

NR NR

(continued)






Peul (2007)30 Peul (2008)95 Lequin (2013)96 High (continued)

Cumulative score on VAS 100 for back pain 0 to 52w SG1: AUC 884.2 (SE 77.6) NS1: AUC 1047.9 (SE 77.6) AMD: -163.7 (95% CI, -379.9 to 52.5) Cumulative score on VAS 100 for back pain 0 to 2y SG1: AUC 1526.3 (SE 169.7) NS1: AUC 1,734.1 (SE 182.3) AMD: -207.8 (95% CI, -702.0 to 286.4) Cumulative score on VAS 100 for back pain 0 to 5 years: SG1: NR NS1: NR Reported as no significant difference SF-36 bodily pain, mean (SD) Baseline: SG1 21.9 (16.6); NS1 23.9 (18.1) AMD (95% CI); [positive AMD favors SG1] 8w: 8.4 (3.2 to 13.5) 26w: 3.3 (-1.8 to 8.4) 52w: 2.7 (-2.6 to 7.9) 2y: -2.3 (-7.3 to 2.7) 5y: NR Sciatica index, Bothersomeness Score, mean (SD) Baseline: SG1 14.6 (5.1); NS1 14.5 (4.1) AMD (95% CI); [negative AMD favors SG1] 8w: -3.6 (-4.9 to -2.3) 26w: -1.2 (-1.3 to -0.1) 52w: -0.4 (-1.5 to 0.7) 2y and 5y: NR

(continued)







Sciatica index, Frequency Score, mean (SD) Baseline: SG1 16.0 (4.6); NS1 16.2 (4.2) AMD (95% CI); [negative AMD favors SG1] 8w: -4.0 (-5.3 to -2.7) 26w: -1.8 (-1.9 to -0.7) 52w: -0.5 (-1.8 to 0.8) 2y and 5y:NR

Ruetten (2008)29 High

SG1: Endoscopic (interlaminar or transforaminal) discectomy N randomized: 100 N analyzed: 91 (91%) SG2: Microdiscectomy N randomized: 100 N analyzed: 87 (87%)

VAS 100 mm leg pain, mean (SD) Baseline: SG1 75 (NR); SG2 71 (NR) 12w: SG1 6 (NR); SG2 9 (NR) Calculated AMD (adjusted for baseline): -7 26w: SG1 9 (NR); SG2 7 (NR) Calculated AMD (adjusted for baseline): -2 52w: SG1 9 (NR); SG2 11 (NR) Calculated AMD (adjusted for baseline): -6 2y: SG1: 8 (NR): SG2 9 (NR) Calculated AMD (adjusted for baseline): -5 Between-group differences reported as NS. VAS 100 mm for back pain, mean (SD) Baseline: SG1 19 (NR); SG2 15 (NR) 12w: SG1 15 (NR); SG2 20 (NR) Calculated AMD (adjusted for baseline): -9 26w: SG1 16 (NR); SG2 22 (NR) Calculated AMD (adjusted for baseline): -10 52w: SG1 17 (NR); SG2 19 (NR) Calculated AMD (adjusted for baseline): -6 2y: SG1 11 (NR); SG2 18 (NR) Calculated AMD (adjusted for baseline): -11 Between-group differences reported as NS.

NASS Neurology Score, mean (SD) Baseline: SG1 31 (NR); SG2 2.9 (NR) 12w: SG1 2 (NR); SG2 2 (NR) 26w: SG1 2.1 (NR); SG2 2.3 (NR) 52w: SG1 1.9 (NR); SG2 1.7 (NR) 2y: SG1 2.1 (NR); SG2 1.9 (NR) Results reported as NS.

NR

(continued)






Ruetten (2008)29 High (continued)

North American Spine Society Pain Score, mean (SD) Baseline SG1 4.6 (NR); SG2 4.2(NR) 12w: SG1 2 (NR); SG2 2.4 (NR) Calculated AMD (adjusted for baseline): -0.8 26w: SG1 2.2 (NR); SG2 2.6 (NR) Calculated AMD (adjusted for baseline): -0.8 52w: SG1 2.2 (NR); SG2 2.4 (NR) Calculated AMD (adjusted for baseline): -0.6 2y: SG1 2.1 (NR); SG2 2.3 (NR) Calculated AMD (adjusted for baseline): -0.6 Between-group differences reported as NS. N (%) with no leg pain at 2y SG1: 77 (85%) SG2: 69 (79%) Results reported as NS. N (%) with leg pain occasionally or pain was greatly reduced at 2y SG1: 12 (13%) SG2: 13 (15%) Results reported as NS. N (%) with no improvement in leg pain at 2y SG1: 2 (2%) SG2: 5 (6%) Results reported as NS. N (%) with progradient back pain at unspecified time point SG1: 2 (2.2%) SG2: 10 (11.5%) P < 0.001

(continued)







SG1: Revision endoscopic discectomy N randomized: 50 N analyzed: 45 (90%) SG2: Revision microdiscectomy N randomized: 50 N analyzed: 42 (84%)

VAS 100 mm leg pain, mean (SD) Baseline: SG1 85 (NR); SG2 79 (NR) 12w: SG1 8 (NR); SG2 12 (NR) 26w: SG1 10 (NR); SG2 12 (NR) 52w: SG1 12 (NR); SG2 9 (NR) 2y: SG1 8 (NR); SG2 10 (NR) Differences between groups reported as NS. VAS 100 mm back pain, mean (SD) Baseline: SG1 14 (NR); SG2 15 (NR) 12w: SG1 14 (NR); SG2 13 (NR) 26w: SG1 12 (NR); SG2 12 (NR) 52w: SG1 16 (NR); SG2 15 (NR) 2y: SG1 15 (NR); SG2 14 (NR) Differences between groups reported as NS. North American Spine Society Pain score, mean (SD) Baseline: SG1 4 (NR); SG2 4 (NR) 12w: SG1 1.9 (NR); SG2 2.1 (NR) 26w: SG1 1.9 (NR); SG2 2.0 (NR) 52w: SG1 2.1 (NR); SG2 2.2 (NR) 2y: SG1 2.1 (NR); SG2 2.1 (NR) Differences between groups reported as NS. N (%) with no leg pain at 2y SG1: 37 (82%) SG2: 32 (76%) Differences between groups reported as NS.

North American Spine Society Neurology Score, mean (SD) Baseline: SG1 3 (NR); SG2 5 (NR) 12w: SG1 2.2 (NR); SG2 2.1 (NR) 26w: SG: 2.0 (NR); SG2 2.1 (NR) 52w: SG1 2.2 (NR); SG2 2.3 (NR) 2y: SG1 2.1 (NR); SG2 2.3 (NR) Differences between groups reported as NS.

(continued)






Ryang (2008)28 Gempt (2013)98 High

SG1: Trocar microdiscectomy N randomized: 30 N analyzed: unclear SG2: Microdiscectomy N randomized: 30 N analyzed: unclear

Outcomes measured over average follow-up of 1.33y (range 26w to 2.17y) Additional long-term outcomes reported for 38 participants over average followup of 2.8y (range 52w to 4.5y) VAS 10 cm Pain Score, mean (SD) Baseline: SG1 6.9 (2.4); SG2 7.3 (2.3), P=0.60 Followup: SG1 2.1 (2.4); SG2 2.1 (2.4), P=0.86 Long-term followup: SG1 median 0.5 (range 0 to 7); SG2 1.65 (0 to 7.5), P reported as NS N (%) with radicular pain Baseline: SG1 27 (90%), SG2 29 (97%), P=0.31 Followup: SG1 1 (3%); SG2 5 (17%), P=0.11 SF-36 bodily pain, mean(SD) Baseline: SG1 22.4 (22.8); SG2 19.1 (17.3), P=0.84 Followup: SG1 68.9 (31.9); SG2 70.2 (25.1), P=0.95 Long-term followup: SG1 58.0 (27.1); 69.4 (26.6), P=0.198

Outcomes measured over average followup of 1.33y (range 26w to 2.17y) N (%) with sensory deficits Baseline: SG1 27 (90%); SG2 22 (73%), P=0.10 Followup: SG1 12 (40%); SG2 13 (43%), P=0.31 N (%) with motor deficits Baseline: SG1 16 (53%); SG2 15 (50%), P=0.61 Followup: SG1 8 (27%); SG2 7 (23%), P=0.86

Outcomes measured over average follow-up of 1.33y (range 26w to 2.17y) Additional long-term outcomes reported for 38 participants over average of 2.8y (range 52w to 4.5y) SF-36 Physical Component Summary, mean (SD) Baseline: SG1 29.3 (7.9); SG2 27.3 (5.9), P=0.44 Followup: SG1 47.6 (10.7), SG2 47.5 (9.4), P=0.79 Long-term followup: SG1 42.6 (10.8); SG2 48.5 (8.7), P=0.081 SF-36 Mental Component Summary, mean (SD) Baseline: SG1 39.5 (12.4); SG2 42.3 (14.8), P=0.51 Followup: SG1 44.0 (13.2), SG2 51.9 (7.8), P=0.03 Long-term followup: SG1 48.4 (9.4); SG2 48.8 (10.5), P=0.892

(continued)






Sasaoka (2006)24 High

SG1: Microendoscopic discectomy N randomized: 15 N analyzed: unclear SG2: Microdiscectomy N randomized: 11 N analyzed: unclear

N (%) with residual low back pain or lumbar discomfort at 52w SG1: NR (36.7%) SG2: NR (66.7%) P NR but difference reported as significant.

NR NR

Teli (2010)27 Some concerns

SG1: Microendoscopic discectomy N randomized: NR N analyzed: 70 SG2: Microdiscectomy N randomized: NR N analyzed: 72 SG3: Discectomy N randomized: NR N analyzed: 70

VAS 10 cm leg pain, mean (SD) Baseline: SG1 8(1); SG2 8 (1); SG3 8 (1) 26w: SG1 2 (1); SG2 2 (1); SG3 2 (1) 52w: SG1 1 (1); SG2 1 (1); SG3 1 (1) 2y: SG1 2 (1); SG2 2 (1); SG3 2 (1) No difference among the three groups at any time point, P=0.73 VAS 10 cm back pain, mean (SD) Baseline: SG1 3 (1); SG2 4 (1); SG3 3 (1) 26w: SG1 2 (1); SG2 2 (1); SG3 1 (1) 52w: SG1 1 (1); SG2 1 (1); SG3 1 (1) 2y: SG1 2 (1); SG2 2 (1); SG3 1 (1) No difference among the three groups at any time point, P=0.75

NR SF-36 Physical Health Component Summary, mean (SD) Baseline: SG1 20 (4); SG2 21 (4); SG3 22 (4) 26w: SG1 42 (4); SG2 42 (4); SG3 42 (4) Calculated AMD 1 52w: SG1 44 (4); SG2 45 (4); SG3 44 (4) Calculated AMD 2 2y: SG1 39 (6); SG2 40 (6); SG3 38 (6) Calculated AMD 3 No difference among the three groups at any time point, P=0.68

(continued)






Teli (2010)27 Some concerns (continued)

SF-36 Mental Health Component Summary Baseline: SG1 22 (3); SG2 21 (2); SG3 23 (2) 26w: SG1 38 (4); SG2 40 (4); SG3 40 (3) Calculated AMD 2 52w: SG1 40 (4); SG2 40 (4); SG3 42 (3) Calculated AMD 0 2y: SG1 38 (5); SG2 39 (6); SG3 39 (3) Calculated AMD 2 No difference among the three groups at any time point, P=0.78

Thome (2005)26 Barth (2008)99 Some concerns

SG1: Sequestrectomy N randomized: 42 N analyzed: 42 (100%) SG2: Microdiscectomy N randomized: 42 N analyzed: 42 (100%)

VAS 10 cm leg pain, mean (SD) Baseline: SG1 5.9 (2.6); SG2 6.7 (2.3) 12w to 26w: SG1 0.7 (1.7); SG2 1.3 (2.5) 52w to 1.5y: SG1 0.6 (1.4); SG2 0.8 (1.7) 2y: SG1 1.2 (1.8); SG2 1.6 (2.4) P>0.05 for difference between groups over time VAS 10 cm back pain, mean (SD) Baseline: SG1 5.2 (2.6); SG2 5.9 (2.5) 12w to 26w: SG1 0.9 (1.4); SG2 1.6 (2.5) 52w to 1.5y: SG1 1.0 (1.7); SG2 1.6 (2.1) 2y: SG1 1.8 (1.9); SG2 2.9 (2.6) P>0.05 for difference between groups over time SF-36 bodily pain, mean (SD) Baseline: SG1 24.4 (13.1); SG2 19.0 (15.0), P=0.09 12w to 26w: SG1 68.6 (19.4); SG2 60.2 (27.6), P=0.14

N (%) of patients with improvement in sensory deficit, 12w to 26w SG1: Exact value NR SG2: Exact value NR P=0.52 N (%) of patients with improvement in motor deficit, 12w to 26w SG1: Exact value NR SG2: Exact value NR P=0.74 Change in sensory index, motor grade, straight leg raise test, and reflex index showed no difference between groups between baseline, 12w to 26w and 2y, P for trend for all parameters reported as > 0.278.

SF-36 Mental Component Summary Score, mean (SD) Baseline: SG1 46.5 (11.8); SG2 47.0 (12.5), P=0.87 12w to 26w: SG1 53.6 (9.8); SG2: 50.6 (12.0), P=0.26 SF-36 Physical Component Summary Score, mean (SD) Baseline: SG1 28.8 (6.6); SG2 28.5 (8.1), P=0.87 12w to 26w: SG1 43.6 (9.7); SG2 41.5 (10.7), P=0.41

(continued)






Thome (2005)26 Barth (2008)99 Some concerns (continued)

2y: SG1 71.6 (25.4); SG2 59.8 N (%) with VAS 10 cm leg pain >3 Baseline: SG1 NR (79), SG2 (NR) 88, P=0.38 12w to 26w: SG1 NR (5%); SG2 NR (18%), P=0.15 52w to 1.5y: SG1 NR (5%); SG2 NR (11%), P=0.43 N (%) with VAS 10 cm back pain >3 Baseline: SG1 NR (74%), SG2 NR (80%), P=0.60 12w to 26w: SG1 NR (5%); SG2 NR (16%), P=0.26 52w to 1.5y: SG1 NR (13%); SG2 NR (19%), P=0.54 (27.7), P=0.064

N (%) of patients reporting specified changes in sensory deficit at 12w to 26w Much better: SG1 NR (73%); SG2 NR (71%) Better: SG1 NR (14%); SG2 NR (18%) Equal: SG1 NR (13%); SG2 NR (11%) Worse: SG1 NR (0%); SG2 NR (0%) Much Worse: SG1 NR (0%); SG2 NR (0%) P=0.969 N (%) of patients reporting specified changes in sensory deficit at 2y Much better: SG1 NR (68%); SG2 NR (54%) Better: SG1 NR (16%); SG2 NR (20%) Equal: SG1 NR (16%); SG2 NR (11%) Worse: SG1 NR (0%); SG2 NR (6%) Much Worse: SG1 NR (0%); SG2 NR (9%) P=0.061 P for trend over time=0.034 N (%) of patients reporting specified changes in motor deficit at 12w to 26w Much better: SG1 NR (67%); SG2 NR (79%) Better: SG1 NR (22%); SG2 NR (15%) Equal: SG1 NR (11%); SG2 NR (3%) Worse: SG1 NR (0%); SG2 NR (3%) Much Worse: SG1 NR (0%); SG2 NR (0%) P=0.390

(continued)







N (%) of patients reporting specified changes in motor deficit at 2y Much better: SG1 NR (84%); SG2 NR (63%) Better: SG1 NR (8%); SG2 NR (23%) Equal: SG1 NR (8%); SG2 NR (6%) Worse: SG1 NR (0%); SG2 NR (0%) Much Worse: SG1 NR (0%); SG2 NR (8%) P=0.041 P for trend over time=0.004

Tullberg (1993)25 Some concerns

SG1: Microdiscectomy N randomized: 30 N analyzed: 29 (97%) SG2: Discectomy N randomized: 30 N analyzed: 29 (97%)

VAS 10 cm leg pain, mean (SD) Baseline: SG1 7.0 (NR); SG2 7.0 (NR) 52w: SG1 2.1 (NR); SG2 2.3 (NR) VAS 10 cm back pain, mean (SD) Baseline: SG1 3.6 (NR); SG2 3.7 (NR) 52w: SG1 1.6 (NR); SG2 1.8 (NR)

NR NR

Weber (1983)5 High

SG1: Discectomy N randomized: 60 N analyzed: 60 (100%) NS1: Conservative management N randomized: 66 N analyzed: 66 (100%)

N (%) in specified category of radiating pain at 4y No pain: SG1 36 (63.2%); NS1 38 (57.6%); Calculated P=0.86 Some pain: SG1 15 (26.3%); NS1 21 (31.8%) Considerable pain: SG1 6 (10.5%); NS1 7 (10.6%) N (%) in specified category of radiating pain at 10y No pain: SG1 43 (84.3%); NS1 52 (78.8%); Calculated P=0.41 Some pain: SG1 8 (14.0%); NS1 14 (21.2%) Considerable pain: SG1 0 (0%); NS1 0 (0%)

NR NR

(continued)






Weinstein (2006) 21 Weinstein (2008)45 Lurie Jon (2014)44 High

SG1: Discectomy N randomized: 245 N analyzed: N=232 in main study's primary analyses. 52w: 202 (82.47%) 2y: 186 (75.9%) 3y: 180 (73.5%) 4y: 149 (60.8%) 8y: 157 (64.1%) NS1: Conservative management N randomized: 256 N analyzed: N=240 included in main study's primary analyses. 52w: 213 (83.2%) 2y: 187 (73.0%) 3y: 170 (66.4%) 4y: 150 (58.6%) 8y: 152 (59.4%)

SF-36 bodily pain, mean (SD) Baseline: SG1 27.1 (18.5); NS1 26.7 (17.4) AMD (95% CI) [positive AMD favors SG1] 12w: 2.9 (-2.2 to 8.0) 52w: 2.8 (-2.3 to 7.8) 2y: 3.2 (-2.0 to 8.4) 4y: 4.5 (-1.2 to 10.3) 8y: 0.7 (-5.2 to 6.6) RM 0 to 2y: AMD NR, P=0.74 RM0 to 4y: AMD NR, P=0.15 RM0 to 8y: AMD NR, P=0.22 Sciatica index, Bothersomeness Score, mean (SD) Baseline: SG1 15.4 (5.1); NS1 15.0 (5.3) AMD (95% CI) [negative AMD favors SG1] 12w: -2.1 (-3.4 to -0.9) 52w: -1.6 (-2.9 to -0.4) 2y: -1.6 (-2.9 to -0.3) 4y: -1.8 (-3.2 to -0.4) 8y: -1.5 (-2.9 to -0.2) RM 0 to 2y: AMD NR, P=0.003 favoring SG1 RM 0 to 4y: AMD NR, P NR for the ITT analysis RM 0 to 8y: AMD NR, P=0.005 favoring SG1

NR NR

Abbreviations: AMD = absolute mean difference; AUC = area under the curve; CI = confidence interval; cm = centimeter; ITT = intention-to-treat; mm = millimeter; N =

number; NR = not reported; NS = not significant; NS1 = nonsurgical intervention group; RM = repeated measures analysis; SD = standard deviation; SE = standard error; SF =

short-form; SG1 = surgical intervention group; VAS = visual analogue scale; w = week(s); y = year(s)

Note: For continuous outcome measures, studies either reported 1) the difference in mean scores at a follow-up time point (e.g. mean score in SG1 minus mean score in NS1 at 6

weeks) or 2) mean change from baseline scores at a follow-up time point (e.g., mean change in score in SG1 minus mean change in score for NS1 at 6 weeks). The absolute mean

difference (AMD) between groups reported or calculated in this table reflects the mean difference between groups with respect to the change in score.

For outcomes where a higher score represents fewer symptoms, a positive AMD means the intervention group (SG1) improves symptoms more than the comparator group

(NS1 or SG2 or SG3).



For outcomes where a lower score represents fewer symptoms, a negative AMD means the inter intervention group (SG1) improves symptoms more than the comparator

group (NS1 or SG2 or SG3).



Table C-4. Individual study findings related to efficacy outcomes Part 2 (functioning/disability, return to work, other efficacy outcomes)


Intervention Names N randomized N analyzed

Physical, Social, or Psychological Functioning and Disability Return to Work Other

Arts (2009)38 Arts (2011)43 Low

SG1: Tubular discectomy N randomized: 167 N analyzed: 166 (99.4%) SG2: Microdiscectomy N randomized: 161 N analyzed: 159 (98.8%)

Roland-Morris Disability Questionnaire, mean (SD) Baseline: SG1 16.0 (4.4); SG2 16.3 (4.3) AMD (95% CI); [Negative AMD favors SG1] 4w: 0.2 (−1.1 to 1.4) 8w: 0.8 (−0.4 to 2.1) 26w: 1.0 (−0.2 to 2.3) 52w: 1.3 (0.03 to 2.6) 2y: 0.8 (-0.5 to 2.1) RM 1w to 52w: 0.8 (−0.2 to 1.7), P=0.11 main treatment effect, P=0.50 treatment X time interaction RM 1w to 2y: 0.6 (-0.3 to 1.6), P=0.17 main treatment effect, P=0.15 treatment X time interaction SF-36 physical functioning, mean (SD) Baseline: SG1 36.7 (20.6); SG2 34.9 (20.7) AMD (95% CI); [Positive AMD favors SG1] 4w: -1.1 (−5.6 to 3.3) 8w: -3.3 (−7.8 to 1.1) 26w: -3.9 (−8.3 to 0.6) 52w: -4.8 (-9.3 to -0.2) 2y: -3.4 (-8.2 to 1.4) RM 1w to 52w: -3.1 (−6.8 to 0.7), P= 0.11 main treatment effect, P=0.27 treatment X time interaction RM 1w to 2y: -2.8 (-6.5 to 0.9), P=0.14 main treatment effect, P=0.33 treatment X time interaction

NR

N (%) of patients with complete or nearly complete recovery, based on 7-pt Likert scale [OR >1 favors SG1] 4w: SG1 NR (62%); SG2 NR (66%); OR 0.84 (95% CI 0.53 to 1.30) 8w: SG1 NR (63%); SG2 NR (75%); OR 0.56 (95% CI 0.35 to 0.92) 26w: SG1 NR (67%); SG2 NR (77%); OR 0.62 (95% CI 0.38 to 1.0) 52w: SG1 NR (69%); SG2 NR (79%), OR 0.59 (95% CI 0.35 to 0.99) 2y: SG NR (71%); SG2 NR (77%); OR 0.76 (95% CI 0.45 to 1.28) Relative difference in rate of complete or nearly complete recovery, based on 7-pt Likert scale, unadjusted HR (95% CI) [HR >1 favors SG1] 1w to 52w: 0.92 (0.73 to 1.17) 1w to 2y: 0.93 (0.74 to 1.17)

(continued)



Table C-4. Individual study findings related to efficacy outcomes Part 2 (functioning/disability, return to work, other efficacy outcomes) (continued)




Prolo Scale, Functional Score, mean (SD) Baseline:SG1 0.8 (0.5); SG2 0.7 (0.5) AMD (95% CI); [Positive AMD favors SG1] 4w: 0 (-0.3 to 0.2) 8w: -0.1 (-0.3 to 0.2) 26w: -0.2 (-0.5 to 0) 52w: -0.2 (-0.5 to 0) 2y: NR RM 1w to 52w: -0.1 (-0.3 to 0), P=0.16 main treatment effect, P=0.43 treatment X time interaction Prolo Scale, Economic Score, mean (SD) Baseline: SG1 1.5 (1.6); SG2 1.3 (1.6) AMD (95% CI); [Positive AMD favors SG1] 4w: 0.2 (−0.1 to 0.5) 8w: 0.1 (−0.2 to 0.4) 26w: 0.1 (−0.2 to 0.4) 52w: 0 (−0.3 to 0.3) 2y: NR RM 1w to 52w: 0.1 (-0.1 to 0.3); P=0.47 main treatment effect, P=0.76 treatment X time interaction

(continued)








SG1: Percutaneous laser disc decompression N randomized: 57 N analyzed: 55 (96.5%) SG2: Microdiscectomy N randomized: 58 N analyzed: 57 (98.3%)

Roland Morris Disability Questionnaire, mean (SD) Baseline: SG1 15.7 (4.9);SG2 15.5 (4.7) AMD (95% CI); [Negative AMD favors SG1] 4w: -2.5 (-4.7 To -0.2) 8w: 0.1 (-2.1 to 2.3)[Primary Outcome] 26w: 2.2 (-0.1 to 4.4) 52w: 1.1(-1.1 to 3.4)[Primary Outcome] 2y: -0.1 (-2.4 to 2.2) RM 1w to 52w: 0.2 (-1.2 to 1.6)[Primary Outcome] RM 1w to 2y: 0.0 (-1.3 to 1.3), P=1.00 main treatment effect, P=0.06 treatment X time interaction SF-36 physical functioning, mean (SD) Baseline: SG1 41.0 (22.6); SG2 38.6 (20.9) AMD (95% CI); [Positive AMD favors SG1] 4w: 18.4 (10.0 to 26.8) 8w: 5.6 (-2.7 to 13.9) 26w: -3.2 (-11.6 to 5.1) 52w: -3.2 (-11.6 to 5.2) 2y: 4.3 (-4.5 to 13.2) RM 1w to 52w: 5.3 (-0.7 to 11.2) RM 1w to 2y: 6.1(0.5 to 11.7), P=0.03 main treatment effect, P=0.001 treatment X time interaction

NR N (%) with complete or nearly complete recovery, based on 7-pt Likert scale [OR > 1 favors SG1] 52w: SG1 NR (69%); SG2 NR (75%); OR 0.81 (95% CI 0.4 to 1.9) 2y: SG1 NR (70.8%); SG2 NR (60.8%); OR 1.6 (95% CI 0.7 to 3.6) Relative difference in time to complete or nearly complete recovery, HR (95% CI) [HR > 1 favors SG1] 52w: 0.64 (0.42 to 0.97) 2y: 0.64 (0.43 to 0.96)

(continued)







Brouwer (2015)37 Brouwer (2017)97 Some concerns (continued)

Prolo Scale, Functioning Score, mean (SD) Baseline: SG1 1.1 (0.6); SG2 0.9 (0.5) AMD (95% CI); [Positive AMD favors SG1] 4w: 0.2 (-0.2 to 0.6) 8w: -0.2 (-0.6 to 0.3) 26w: 0.1 (-0.3 to 0.5) 52w: -0.2 (-0.5 to 0.2) RM 1w to 52w: 0.0 (-0.3 to 0.3) Prolo Scale, Economic Score, mean (SD) Baseline: SG1 1.7 (1.7); SG2 2.1 (1.7) AMD (95% CI); [Positive AMD favors SG1] 4w: 1.1 (0.5 to 1.6) 8w: 0.2 (-0.3 to 0.8) 26w: -0.7 (-1.3 to 0.2) 52w: 0.1 (-0.4 to 0.7) RM 1w to 52w: 0.3 (-0.1 to 0.7)


SG1: Automated percutaneous lumbar discectomy (APLD) N randomized: 31 N analyzed: 31 (100%) SG2: Microdiscectomy N randomized: 40 N analyzed: 40 (100%)

NR Returned to work or previous level of activity by 12w SG1: NR SG2: 37 (92.5%)

N (%) with excellent/good outcome based on MacNab criteria SG1: 9 (29%) SG2: 32 (80%) P < 0.001

(continued)








SG1: Percutaneous disc decompression N randomized: 31 N analyzed: 31 (100%) NS1: Conservative management N randomized: 31 N analyzed: 31 (100%)

NR N (%) reporting that pain affected their occupational status 12w: SG1 4 (12.9%); NS1 3 (9.7%) Calculated P=1.0 52w: SG1 4 (12.9%); NS1 22 (71%) Calculated P<0.001 2y: SG1 4 (12.9%); NS1 22 (71%) Calculated P<0.001

NR


SG1: Microscopically assisted percutaneous nucleotomy N randomized: 52 N analyzed: 52 (100%) SG2: Microdiscectomy N randomized: 48 N analyzed: 48 (100%)

Oswestry disability index (RM at 52w) Significant within group improvement over time in both groups (P< 0.001) No between group difference, P=0.08

Duration of postoperative inability to work, mean (SD) in weeks Overall: 7 (NR), NR by group

NR

(continued)







Gerszten (2010)39 Some concerns (6w outcomes) High (12w and 26w outcomes)

SG1: Plasma disc decompression with coblation technology (PDD) N randomized: 46 N analyzed: 29 (64% of ITT sample) at 26w NS1: Epidural steroid injection (TFESI) N randomized: 44 N analyzed: 28 (70% of ITT sample) at 26w

Oswestry disability index, mean (SD) Baseline: SG1: 42 (14); NS1 43 (17) Change in score, mean (SE); [larger negative change favors SG1] 6w: SG1 -13 (3); NS1 -5 (2); P=0.002 12w: SG1 -11 (3); NS1 4 (2); P= 0.002 26w: SG1 -14 (4); NS1 -4 (2); P=0.002 SF-36 physical functioning at 26w Larger improvements in SG1 compared to NS1, P=0.0016 SF-36 Social Functioning at 26w Larger improvements in SG1 compared to NS1, P=0.0312 SF-36 Physical Component Summary at 26w Larger improvements in SG1 compared to NS1, P=0.0040 No significant difference between groups for the following SF-36 domains: role physical, general health, vitality, role emotional, mental health, or mental components summary

N (%) of participants working full or part time at 26w Reported as similar in both groups (69% to 70%)

NR

(continued)








SG1: Automated percutaneous discectomy, endoscopic percutaneous discectomy (APD/EPD) N randomized: 21 N analyzed: 17 (81.0%) at 26w SG2: Discectomy N randomized: 13 N analyzed: 10 (76.9%) at 26w

Modified Roland Disability Score at 26w, mean (SD) Baseline: SG1 16.9 (4.9) SG2 17.3 (4.1) 26w: SG1 6.12 (7.2) SG2 6.5 (6.1) (P=0.74) Calculated AMD (adjusted for baseline) 0.02 SF-36 physical functioning at 26w, mean (SD) Baseline: SG1 36.0 (27.1); SG2 37.2 (15.8) 26w: SG1 74.7 (27.6) SG2 73.0 (15.7) (P=0.96) Calculated AMD (adjusted for baseline) 2.9

NR

N (%) with success at 26w SG1: 7 (41.2%) SG2: 4 (40%) P=0.95 Success defined as either excellent or good on author defined outcome assessment matrix incorporating 4 dimensions (pain frequency and severity, ability to participate in work activities, ability to participate in leisure activities, and analgesic use N (%) with success at 26w, based on MacNab Criteria SG1: 11 (64.7%) SG2: 6 (60%) P=0.81



NR NR NR

(continued)









NR Duration of postoperative disability in time lost from work or until able to resume normal activity, mean (SD) in weeks at unspecified follow-up time SG1: 3.9 (NR) SG2: 7 (NR)

Follow-up time for these measures NR N (%) very satisfied with operative result based on self-report SG1: 22 (73%) SG2: 20 (67%) Calculated P=0.78 N (%) with satisfactory outcome SG1: 29 (97%) SG2: 28 (93%) Calculated P=1.0 Satisfactory outcome defined as either an excellent (radicular symptoms had ceased, the tension signs had become negative, the patient had returned to his or her previous occupation or to normal activity, and the patient expressed satisfaction with the results of the operative procedure) or good (if the above criteria were met but the patient had residual back pain and had had to modify his or her occupation) outcome)



NR NR N (%) with excellent/good outcome, based on MacNab criteria, follow-up time point unspecified SG1: 9 (90%) SG2: 11 (91.6%) Calculated P=1.0

(continued)







Mayer (1993)32 High


NR Duration of postoperative disability, mean (range) in weeks SG1: 7.7 (1 to 26) SG2: 22.9 (4 to 52) N (%) returning to work SG1: 19 (95%) SG2: 13 (65%) Calculated P=0.044

Clinical score, mean (SD) Baseline: SG1 4.55 (0.99); SG2 4.2 (0.98) 2y: SG1 8.23 (1.3); SG2 7.67 (1.9) P <0.005, favoring SG1 Clinical scoring system (modified from the system of Suezawa and Schreiber, based on pain, sensory, motor, and reflexes. A total score of 9 to 10 indicates an excellent condition, 7 to 8 good, 6 to 7 moderate, 5 or less poor. N (%) with specified clinical score at 2y Excellent: SG1 13 (68.4%); SG2 7 (35%) Good: SG1 4 (21%); SG2 6 (30%) Moderate: SG1 3 (16%); SG2 4 (20%) Bad: SG1 0 (0%); SG2 3 (15%) N (%) with specified self-reported success of surgery at 2y Excellent: SG1 9 (47%); SG2 8 (40%) Good: SG1 5 (26%); SG2 3 (15%) Satisfied: SG1 6 (32%); SG2 6 (30%) Bad: SG1 0 (0%); SG2 0 (0%)

(continued)









Roland Score, mean (SD) Baseline: SG1 10.1 (5.7); NS1 12.0 (5.4) 6w: SG1 9.4 (6.4); NS1 9.5 (6.0) 12w: SG1 7.2 (6.9); NS1 9.0 (6.2) RM 0w to 12w: AMD NR, P=0.033 for time effect, P=0.199 for main treatment effect, P=0.760 for treatment X time interaction SF-36 physical functioning, mean (SD) Baseline: SG1 42.7 (22.7); NS1 47.4 (24.8) 6w: SG1 51.3 (28.2); NS1 54.8 (24.4) Calculated AMD 1.2 (adj. for baseline) 12w: SG1 65.8 (27.6); NS1 59.0 (25.4) Calculated AMD 11.5 (adj. for baseline) RM 0 to 12w (RM): AMD NR, P=0.034 for time effect, P=0.720 for main treatment effect, P=0.448 for treatment X time interaction SF-36 role physical, mean (SD) Baseline: SG1: 17.5 (32.5); NS1 18.8 (26.7) 6w: SG1 15.0 (33.8); NS1 26.3 (37.6) 12w: SG1 28.8 (37.4); NS1 32.5 (38.1) RM 0w to 12w (RM): AMD NR, P=0.126 for time effect, P=0.719 for main treatment effect, P=0.038 for treatment X time interaction SF-36 role emotional, mean (SD) Baseline: SG1 60.8 (41.0); NS1 53.4 (50.0) 6w: SG1 63.3 (47.0); NS1 66.7 (45.7) 12w: SG1 65.0 (43.9); NS1 74.5 (36.4) RM 0w to 12w: AMD NR, P=0.034 for time effect, P=0.715 for main treatment effect, P=0.410 for treatment X time interaction

NR NR

(continued)







McMorland (2010)22 Some concerns (continued)

SF-36 social function, mean (SD) Baseline: SG1 50.2 (29.0); NS1 52.9 (33.0) 6w: SG1 54.1 (30.9); NS1 62.7 (20.0) 12w: SG1 67.3 (34.7); NS1 73.6 (19.7) RM 0w to 12w: AMD NR, P=0.138 for time effect, P=0.938 for main treatment effect, P=0.596 for treatment X time interaction SF-36 mental health, mean (SD) Baseline: SG1 69.2 (13.0); NS1 69.0 (21.3) 6w: SG1 77.7 (16.8); NS1 78.6 (11.6) 12w: SG1 83.2 (10.6); NS1 82.8 (8.7) RM 0w to 12w: AMD NR, P=0.001 for time effect, P=0.905 for main treatment effect, P=0.990 for treatment X time interaction

North (2005)47 High


Impairment from pain in performing everyday activities Reported as higher in SG1 compared with NS1 but actual values are NR and differences reported as NS Everyday activities defined as work, walk, climb stairs, sleep, engage in sex, drive a car, sit at a table to eat

Reported as no significant treatment differences, but actual values NR

N (%) with successful treatment SG1: 3 (12%) NS1: 9 (47%) P < 0.01 Success defined as at least 50% pain relief and patient satisfaction with treatment. Long-term followup occurred at a mean of 2.9y (range 1.8 to 5.7)

(continued)









Oswestry disability index, mean (SD) Baseline: SG1 39 (15); NS1 39 (14) 6w: SG1 16 (16); NS1 22 (16) Calculated AMD -6 (adj. for baseline) 12w: SG1 8 (11); NS1 14 (14) Calculated AMD -6 (adj. for baseline) 26w: SG1 8 (12); NS1 12 (15) Calculated AMD -4 (adj. for baseline) 52w: SG1 10 (13); NS1 11 (14) Calculated AMD -1 (adj. for baseline) 2y: SG1 6 (9); NS1 11 (16) Calculated AMD -5 (adj. for baseline) RM 0 to 2y, AMD (95% CI) [negative AMD favors SG1] -3 (-10 to 4)

VAS100 work ability, mean (SD) 6w: SG1 68 (27); NS1 63 (32) Calculated AMD: 5 (95% CI, -11.5 to 21.5) 12w: SG1 84 (14); NS1 70 (31) Calculated AMD: 14 (95% CI, 0.60 to 27.4) 26w: SG1 87 (18); NS1 75 (30) Calculated AMD: 12 (95% CI, -2.1 to 26.1) 52w: SG1 82 (26); NS1 81 (27) Calculated AMD: 1 (95% CI, -15.7 to 17.7) 2y: SG1 89 (16); NS1 79 (28) Calculated AMD: 10 (95% CI, -2.8 to 22.8) RM 0 to 2y, AMD (95% CI) [positive AMD favors SG1] 5 (-7 to 18)

N (%) reporting full recovery 6w: SG1 5 (19.2%); NS1 0 (0%), P < 0.05 12w: SG1 5 (19.2%); NS1 4 (15.4%), Calculated P=1.0 26w: NR 52w: SG1 7 (33.3%); NS1 5 (25%), Calculated P=0.73 2y: NR


SG1: Microdiscectomy N randomized: 141 N analyzed: 52w: 140 (99.3%) 2y: 130 (92.2%) 5y: 115 (81.6%) NS1: Conservative management N randomized: 142

Roland Disability Questionnaire, mean (SD) Baseline: SG1 16.5 (4.4); NS1 16.3 (3.9) AMD (95% CI) [negative AMD favors SG1] 8w: -3.1 (-4.3 to -1.7) 26w: -0.8 (-2.1 to 0.5) 52w: -0.4 (-1.7 to 0.9) 2y: -0.5 (-1.8 to 0.8) 5y: 0.1 (-1.3 to 1.4)

NR

Median time to recovery, weeks SG1: 4.0 (95% CI, 3.7 to 4.4) NS1: 12.1 (95% CI, 9.5 to 14.9) AMD NR, P<0.001 Relative difference in time to complete or nearly complete recovery at 52w, HR (95%CI); [HR >1 favors SG1] 1.97 (95% CI, 1.72 to 2.22)

(continued)








N analyzed: 52w: 141 (99.3%) 2y: 130 (91.5%) 5y: 116 (81.7%)

Cumulative score on Roland 0 to 52w SG1: AUC 273.9 (SE 20.7) NS1: AUC 316.3 (SE 18.8) AMD: -42.5 (95% CI, -97.4 to 12.4) Cumulative score on Roland 0 to 2 y SG1: AUC 397.2 (SE 39.0) NS1: AUC 458.2 (SE 35.3) AMD: -61.0 (95% CI, -164.5 to 42.5) SF-36 physical functioning, mean (SD) Baseline: SG1 33.9 (19.6); NS1 34.6 (19.0) AMD (95% CI); [positive AMD favors SG1] 8w: 9.3 (4.4 to 14.2) 26w: 1.5 (-3.4 to 6.4) 52w: 2.2 ( -2.8 to 7.2) Prolo Scale, Functional Score, mean (SD) Baseline: NR AMD (95% CI); [positive AMD favors SG1] 8w: 0.8 (-0.6 to 1.1) 26w: 0.5 (0.2 to 0.7) 52w: -0.04 (-0.3 to 0.2) 2y: NR 5y: NR Prolo Scale, Economic Score, mean (SD) Baseline: NR AMD (95% CI); [positive AMD favors SG1] 8w: -0.5 (-0.8 to -0.1) 26w: 0.1 (-0.3 to 0.5) 52w: -0.2 (-0.6 to 0.2) 2y: NR 5y: NR

Likert scale (7-pt) global perception of recovery, mean score (SE) [negative AMD favors SG1] 8w: SG1 2.2 (0.1); NS1 3.1 (0.1); AMD -0.9 (95% CI, -1.2 to -0.6) 26w: SG1 2.1 (0.1); NS1 2.3 (0.1); AMD -0.2 (95% CI, -0.5 to 0.1) 52w: SG1 1.9 (0.1); NS1 2.1 (0.1); AMD -0.2 (95% CI, -0.4 to 0.1) N (%) reporting complete or nearly complete recovery, based on 7-pt Likert scale at 2y SG1: NR (81.3%) NS1: NR (83.6%) AMD: -2.4% (95% CI, -12.0% to 7.2)

(continued)








SF-36 social functioning, mean (SD) Baseline: SG1 44.6 (30.1); NS1 43.4 (27.1) AMD (95% CI); [positive AMD favors SG1] 8w: 2.3 (-3.7 to 8.3) 26w: 4.5 (-1.4 to 10.6) 52w: 1.3 (-4.7 to 7.3) SF-36 role emotional, mean (SD) Baseline: SG1 51.0 (46.0): NS1 52.4 (46.0) AMD (95% CI); [positive AMD favors SG1] 8w: 3.1 (-3.0 to 9.3) 26w: 3.9 (-2.3 to 10.1) 52w: -1.4 (-7.6 to 4.8)



Oswestry disability index, mean (SD) Baseline: SG1 75 (NR); SG2 73 (NR) 12w: SG1 22 (NR); SG2 26 (NR) Calculated AMD (adjusted for baseline): -6 26w: SG1 21 (NR); SG2 24 (NR) Calculated AMD (adjusted for baseline): -5 52w: SG1 19 (NR); SG2 23 (NR) Calculated AMD (adjusted for baseline): -6 2y: SG1 20 (NR); SG2 24 (NR) Calculated AMD (adjusted for baseline): -6 Between-group differences reported as NS.

Duration of postoperative work disability, mean (SD) in weeks SG1: 3.57 (NR) SG2: 7 (NR) P < 0.01

NR

(continued)









Oswestry disability index, mean (SD) Baseline: SG1 80 (NR); SG2 84 (NR) 12w: SG1 22 (NR); SG2 18 (NR) 26w: SG1 24 (NR); SG2 19 (NR) 52w: SG1 18 (NR); SG2 23 (NR) 2y: SG1 20 (NR); SG2 21 (NR) Differences between groups reported as NS.

Postoperative work disability, mean (SD) in weeks SG1: 4 (NR) SG2: 7.4 (NR) P < 0.01

N (%) satisfied with surgery and would undergo the operation again SG1: 43 (95%) SG2: 36 (86%) Calculated P=0.15



Outcomes measured over average followup of 1.33y (range 26w to 2.17y) Additional long-term outcomes reported for 38 participants over average of 2.8y (range 52w to 4.5y) Oswestry disability index, mean (SD) Baseline: SG1 53.1 (19.2); SG2 56.7 (23.1), P=0.48 Followup: SG1 12.0 (14.0); SG2 12 (18.8), P=0.83 Calculated AMD (adjusted for baseline): 3.6 Long-term followup: SG1 12.95 (11.2); SG2 18.53 (15.37), P reported as NS Calculated AMD (adjusted for baseline): -1.98 SF-36 physical functioning, mean (SD) Baseline: SG1 33 (23.6); SG2 28.5 (26.1), P=0.45 Followup: SG1 74.8 (23.3); SG2 80.4 (19.6), P=0.64, calculated AMD (adjusted for baseline) -10.1 Long-term followup: SG1 72.4 (25.0); SG2 78.2 (20), P=0.436, calculated AMD (adjusted for baseline) 6

NR Long-term outcomes reported for 38 participants over average of 2.8y (range 52w to 4.5y) VAS10 for improvement from baseline to long-term followup (95%CI) SG1: 4.92 (3.23; 6.61); P<0.001 SG2: 4.64 (3.03; 6.26); P<0.001 Between group P value NR.

(continued)







Ryang (2008)28 Gempt (2013)98 High (continued)

SF-36 Role physical, mean (SD) Baseline: SG1 11 (20.5); SG2 14.5 (28.7), P=0.87 Followup: SG1 66.3 (41); SG2 73.2 (37.3), P=0.55 Long-term followup: SG1 58.3 (42.0), SG2 81.9 (36.2), P= 0.080 SF-36 Social Functioning, mean (SD) Baseline: SG1 44 (32.5); SG2 46.5 (30), P=0.67 Followup: SG1 78.3 (27.2); SG2: 88.4 (18), P=0.18 Long-term followup: SG1 80.9 (26.1); SG2 88.2 (20.6), P=0.350 SF-36 Role emotional, mean (SD) Baseline: SG1 38.7 (45.8); SG2 43 (48.1), P=0.8 Followup: SG1 60.6 (46.7); SG2 85.2 (31.1), P=0.03 Long-term followup: SG1 77.8 (37.9); SG2 77.8 (42.8), P=1.00



NR NR Mean (SD) percentage improvement of the Japanese Orthopaedic Association (JOA) score at 52w SG1: 84.7% (NR) SG2: 88.6% (NR) Difference reported as not significant The JOA score comprises of back pain symptoms, leg pain and/or tingling, gait, clinical signs (straight leg-raising test, sensory disturbance, motor disturbance), restriction of activities of daily living, and urinary bladder function

(continued)









Oswestry disability index, mean (SD) Baseline: SG1 40 (4); SG2 41 (4); SG3 39(4) 26w: SG1 12 (4); SG2 12 (4); SG3 12 (4) Calculated AMD (adjusted for baseline) SG1/SG2 1 Calculated AMD (adjusted for baseline) SG2/SG3 2 52w: SG1 14 (4); SG2 13 (4); SG3 13 (4) Calculated AMD (adjusted for baseline) SG1/SG2 2 Calculated AMD (adjusted for baseline) SG2/SG3 2 2y: SG1 14 (6); SG2 16 (5); SG3 15 (3) Calculated AMD (adjusted for baseline) SG1/SG2 1 Calculated AMD (adjusted for baseline) SG2/SG3 1 No difference among the three groups at any time point, P=0.81

NR NR



For all SF-36 scores, higher values represent more favorable outcomes. SF-36 physical functioning, mean (SD) Baseline: SG1 37.9 (21.8); SG2 34.4 (26.9), P=0.53 12w to 26w: SG1 75.1 (24.0); SG2 69.4 (25.6), P=0.32 Calculated AMD 2.2 2y: SG1 82.4 (20.9); SG2 71.9 (23.2), P=0.026 Calculated AMD 7

N (%) reporting specified categories of impairment of work at 12w to 26w Much better: SG1 NR (31%); SG2 NR (33%) Better: SG1 NR (33%); SG2 NR (19%) Equal: SG1 NR (25%); SG2 NR (27%) Worse: SG1 NR (8%); SG2 NR (15%)

N (%) with specified patient satisfaction index scores at 12w to 26w [Lower scores represent more favorable outcome] 1: SG1 NR (67%); SG2 NR (67%) 2: SG1 NR (30%); SG2 NR (24%) 3: SG1 NR (0%); SG2 NR (6%) 4: SG1 NR (3%); SG2 NR (3%) P=0.693

(continued)








SF-36 role physical, mean (SD) Baseline: SG1 14.7 (29.0); SG2 16.4 (29.0), P=0.79 12w to 26w: SG1 53.21 (42.2); SG2 49.5 (43.8), P=0.72 2y: SG1 67.1(44.7); SG2 55.0 (45.7), P=0.249 SF-36 social functioning, mean (SD) Baseline: SG1 54.0(26.6); SG2 56.8 (26.6), P=0.64 12w to 26w: SG1 83.6 (23.8); SG2 85.1 (20.2), P=0.77 2y: SG1 89.8 (18.6); SG2 79.6 (22.5), P=0.013 SF-36 role emotional, mean (SD) Baseline: SG1 58.8 (46.8); SG2 57.1 (48.2), P=0.88 12w to 26w: SG1 82.4 (36.1); SG2 79.8 (36.0), P=0.76 2y: SG1 85.9 (32.5); SG2 73.56 (44.8), P=0.311 Note: the Prolo Scale used in this study was modified from original version, it used a 0 to 5 Likert scale for each subscale, and summed to obtain a total score that ranged from 2 to 10. N (%) with total Prolo score >=7, 12w to 26w [higher proportion reflects more favorable outcome] SG1: NR (92%) SG2: NR (76%) P= 0.11

Much worse: SG1 NR (3%); SG2 NR (7%) P=0.415 N (%) reporting specified categories of impairment of work at 2y Much better: SG1 NR (37%); SG2 NR (31%) Better: SG1 NR (37%); SG2 NR (11%) Equal: SG1 NR (12%); SG2 NR (41%) Worse: SG1 NR (11%); SG2 NR (11%) Much worse: SG1 NR (3%); SG2 NR (6%) P=0.112

N (%) with specified patient satisfaction index scores at 2y [Lower scores represent more favorable outcome] 1: SG1 NR (68%); SG2 NR (56%) 2: SG1 NR (29%); SG2 NR (29%) 3: SG1 NR (3%); SG2 NR (9%) 4: SG1 NR (0%); SG2 NR (6%) P=0.087 N (%) unsatisfied with surgery as measured by patient satisfaction index score of 3 or 4, 12w to 26w SG1: NR (3%) SG2: NR (18%) P = 0.06

(continued)








N (%) with specified total Prolo scores at 12w to 26w 1-4: SG1 NR (3%); SG2 NR (9%) 5-6: SG1 NR (6%); SG2 NR (9%) 7-8: SG1 NR (31%); SG2 NR (12%) 9-10: SG1 NR (60%); SG2 NR (70%) P=0.852 N (%) with specified Prolo scores at 2y 1-4: SG1 NR (0%); SG2 NR (3%) 5-6: SG1 NR (3%); SG2 NR (12%) 7-8: SG1 NR (23%); SG2 NR (19%) 9-10: SG1 NR (74%); SG2 NR (66%) P=0.20



NR

Postoperative sick leave, mean (SD) in weeks Full time sick leave SG1: 10.4 (NR) SG2: 10.1 (NR) Half time sick leave SG1: 2.6 (NR) SG2: 2.9 (NR) N (%) of patients out of work at unspecified time point during followup SG1: 5 (16.7%) SG2: 2 (6.7%) Calculated P=0.42

N (%) with specified opinion on recovery at 52w SG1 Totally recovered: SG1: 11 (37.9%); SG2: 6 (20.7%); Calculated P=0.25 Almost recovered: SG1: 8 (27.6%); SG2: 14 (48.3%); Calculated P=0.18 Good: SG1: 6 (20.7%); SG2: 6 (20.7%); Calculated P=1.0 Unchanged: SG1: 4 (13.8%); SG2: 2 (6.9%); Calculated P=0.67 Worse: SG1: 0 (0%); SG2: 1 (3.4%); Calculated P=1.0

(continued)







Weber (1983)5 High


NR

Cumulative N (%) with permanent incapacitation and receiving disablement benefit 0-4y: SG1 3 (5%); NS1 8 (12.1%); Calculated P=0.21 5-10y: SG1 7 (11.7%); NS1 8 (12.1%); Calculated P=1.0

Result (good, fair, poor, or bad) based on author's evaluation of neurological deficits, working capacity, pain, and mobility of lumbar spine. N (%) with specified result at 52w Good: SG1 39 (65.0%); NS1 24 (36.4%) Fair: SG1 16 (26.7%); NS1 28 (42.4%) Poor: SG1 5 (8.3%); NS1 13 (19.7%) Bad: SG1 0 (0%); NS1 1 (1.5%) P= 0.0015 N (%) with specified result at 4y Good: SG1 40 (70.2%); NS1 34 (51.5%) Fair: SG1 9 (15.8%); NS1 24 (36.4%) Poor: SG1 8 (14.0%); NS1 5 (7.6%) Bad: SG1 0 (0%); NS1 3 (4.5%) Not examined: SG1 2 (3.3); NS1 0 (0%) P reported as NS N (%) with specified result at 10y Good: SG1 35 (63.6%); NS1 37 (56.0%) Fair: SG1 16 (29.1%); NS1 25 (37.9%) Poor: SG1 4 (7.3%); NS1 4 (6.1%) Bad: SG1 0 (0%); NS1 0 (0%) Not examined: SG1 2 (3.3); NS1 0 (0%) P reported as NS Relapses 0 to 4y: SG1 8 (13.3%); NS1 14 (6.1%) 5 to 10y: SG1 13 (21.7%); NS1 11 (16.7%) (5 patients in this period also had a relapse in the prior period in both groups)

(continued)







Weinstein (2006)21 Weinstein (2008)45 Lurie Jon (2014)44 High

SG1: Discectomy /microdiscectomy N randomized: 245 N analyzed: N=232 in main study's primary analyses. 52w: 202 (82.47%) 2y: 186 (75.9%) 3y: 180 (73.5%) 4y: 149 (60.8%) 8y: 157 (64.1%) NS1: Conservative management N randomized: 256 N analyzed: N=240 included in main study's primary analyses. 52w: 213 (83.2%) 2y: 187 (73.0%) 3y: 170 (66.4%) 4y: 150 (58.6%) 8y: 152 (59.4%)

SF-36 physical functioning, mean (SD) Baseline: SG1 39.7 (24.9); NS1 39.2 (25.7) AMD (95% CI) [positive AMD favors SG1] 12w: 2.8 (-2.5 to 8.1) 52w: 1.2 (-4.1 to 6.5) 2y: 0 (-5.4 to 5.5) 4y: 2.2 (-3.7 to 8) 8y: 1.7 (-4.0 to 7.4) RM 0 to 2y: AMD NR, P=0.71 RM 0 to 4y: AMD NR, P=0.42 RM 0 to 8y: AMD NR, P=0.47 Oswestry disability index, mean (SD) Baseline: SG1 47.5 (21.4); NS1 46.3 (20.6) AMD (95% CI) [negative AMD favors SG1] 12w: -4.7 (-9.3 to -0.2) 52w: -3.2 (-7.8 to 1.3) 2y: -2.7 (-7.4 to 1.9) 4y: -3.6 (-8.6 to 1.4) 8y: -4.2 (-9.0 to 0.7) RM 0 to 2y: AMD NR, P=0.21 RM 0 to 4y: AMD NR, P=0.074 RM 0 to 8y: AMD NR, P=0.096

N (%, (SE %) working full or part-time Baseline: SG1 142 (61.2% (NR)); NS1 148 (61.7% (NR)) 12w: SG1 NR, 63.8% (3.3%); NS1 NR, 69.4% (3.1%); AMD (95% CI): -5.6% (-14.5% to 3.4%) 52w: SG1 NR, 76.4% (2.9%); NS1 NR 77.0% (2.8%); AMD (95% CI): -0.6% (-8.6% to 7.3%) 2y: SG1 NR, 74.2% (3.1%); NS1 NR, 76.4% (3.0%); AMD (95% CI): -2.2% (-10.6% to 6.2%) 4y: SG1 71.4% (NR); NS1 75.1% (NR); AMD (95% CI): −3.8 (−13.3, 5.8)

Very/somewhat satisfied with symptoms, mean % (SE); [positive AMD favors SG1] 12w: SG1 54.3% (3.5%); NS1 43.0% (3.4%) AMD:11.3% (95% CI, 1.6% to 20.9%) 52w: SG1 64.7% (3.4%); NS1 58.5% (3.4%) AMD: 6.1% (95% CI, -3.5% to 15.5%) 2y: SG1 68.3% (3.4%); NS1 64.4% (3.5%) AMD: 4.0% (95% CI, -5.6% to 13.5%) 4y: SG1 64.7% (NR); NS1 61.3% (NR) AMD: 3.4% (95% CI, -7.7% to 14.6%) 8y: SG1 74.3% (NR); NS1 67.4% (NR) AMD: 6.8% (95% CI, -3.4% to 17%) RM 0 to 8y: AMD NR, P=0.013 favoring SG1 RM 0 to 4y: AMD NR, P reported as not significant RM 0 to 2y: AMD NR, P=0.17 % (SE %) Self-rated progress: major improvement; [positive AMD favors SG1] 12w: SG1 66.3% (3.3%); NS1 62.1% (3.4%)

(continued)







Weinstein (2006)21 Weinstein (2008)45 Lurie Jon (2014)44 High (continued)

AMD: 4.2% (95% CI, -5.1% to 13.5%) 52w: SG1 75.7% (3.0%); NS1 66.7% (3.2%) AMD: 9.0% (95% CI, 0.3% to 17.6%) 2y: SG1 76.3% (3.1%); NS1 69.3% (3.3%) AMD: 7.0% (95% CI, -1.9% to 15.9%) 4y: SG1 72.5% (NR); NS1: 65% (NR) AMD: 7.5% (95% CI, -3.2% to 18.1%) 8Y: SG1 62.3%; NS1: 58.2% AMD: 4.1% (95% CI, -7% to 15.3%) RM 0 to 2y: AMD NR, P=0.04 favoring SG1 RM 0 to 4y: AMD NR, P reported as NS RM 0 to 8y: AMD NR, P= 0.013 favoring SG1

Abbreviations: AMD = absolute mean difference; AUC = area under the curve; CI = confidence interval; N = number; NR = not reported; NS = not significant; NS1 =

nonsurgical intervention group; RM = repeated measures analysis; HR = hazard ratio; SD = standard deviation; SE = standard error; SF = short-form; SG1 = surgical intervention

group; VAS = visual analogue scale; w = week(s); y = year(s)

Note: For continuous outcome measures, studies either reported 1) the difference in mean scores at a follow-up time point (e.g. mean score in SG1 minus mean score in NS1 at 6

weeks) or 2) mean change from baseline scores at a follow-up time point (e.g., mean change in score in SG1 minus mean change in score for NS1 at 6 weeks). The absolute mean

difference (AMD) between groups reported or calculated in this table reflects the mean difference between groups with respect to the change in score.

For outcomes where a higher score represents fewer symptoms, a positive AMD means the intervention group (SG1) improves symptoms more than the comparator group

(NS1 or SG2 or SG3).

For outcomes where a lower score represents fewer symptoms, a negative AMD means the inter intervention group (SG1) improves symptoms more than the comparator

group (NS1 or SG2 or SG3).



Table C-5. Individual study findings related to safety outcomes

Main Study Author (Year); Follow-up Studies Author (Year); Risk of Bias

Intervention Names N Randomized N Analyzed Mortality Surgical morbidity Reoperations

Persistent Opioid Use

Arts (2009)38 Arts (2011)43 Low

SG1: Tubular discectomy N randomized: 167 N analyzed: 166 (99.4%) SG2: Microdiscectomy N randomized: 161 N analyzed: 159 (98.8%)

All-cause mortality NR Surgical mortality NR

N (%) with intraoperative complications SG1: 20 (12%) SG2: 13 (8%) P=0.27 Dural tear was most common complication in both groups (Ns NR), but no difference between groups (P=0.18). N (%) with postoperative complications SG1: 19 (11%) SG2: 14 (9%) P=0.47

N (%) with reoperations at 52w SG1: 17 (10%) SG2: 11 (7%) P=0.33 Reason for reoperations was recurrent disc herniation in 20 (71%) of cases. N (%) with repeat surgery at 2y SG1: 23 (15%) SG2: 14 (10%) P=0.22 Reason for repeated surgery was recurrent disc herniation (same level) in 25 (68%) of cases

NR


SG1: Percutaneous laser disc decompression N randomized: 57 N analyzed: 55 (96.5%) SG2: Microdiscectomy N randomized: 58 N analyzed: 57 (98.3%)


N (%) with surgical complications SG1: 3 (5%) (all transient nerve root injury) SG2: 6 (11%) (3 CSF leak, 1 micturition problem requiring catheter, 1 transient nerve root injury, 1 surgery at wrong level) Calculated P=0.49 N (%) with technical failure SG1: 5 (9%) SG2: NA

N (%) with reoperations at 52w SG1: 24 (44%) SG2: 9 (16%) Calculated P=0.002

NR

(continued)



Table C-5. Individual study findings related to safety outcomes (continued)





SG1: Automated percutaneous lumbar discectomy N randomized: 31 N analyzed: 31 (100%) SG2: Microdiscectomy N randomized: 40 N analyzed: 40 (100%)


NR N (%) with reoperations at unspecified follow-up time point SG1: 20 (64.5%) (offered microdiscectomy) SG2: 1 (2.5%) (for recurrent disc protrusion) Calculated P<0.001

NR


SG1: Percutaneous disc decompression N randomized: 31 N analyzed: 31 (100%) NS1: Conservative management N randomized: 31 N analyzed: 31 (100%)


N (%) with intraoperative or postoperative complications up to 2y SG1: 0 (0%) NS1: NA

N (%) with reoperations at unspecified follow-up time SG1: 1 (3.2%) NS1: NA

NR

(continued)








SG1: Microscopically-assisted percutaneous nucleotomy N randomized: 52 N analyzed: 52 (100%) SG2: Microdiscectomy N randomized: 48 N analyzed: 48 (100%)


N (%) with intraoperative dural tear SG1: 2 (3.9%) SG2: 3 (6.3%) Calculated P=0.67

N (%) with reoperations SG1: 2 (3.9%) SG2: 5 (10.4%) Calculated P=0.26 N (%) with reoperations due to relapse (same level, same side) Overall: 5 (5%) SG1: 1 (1%) SG2: 4 (4%) Calculated P=0.19 N (%) with reoperations due to progressive disc degeneration with segmental instability Overall: 2 (2%) SG1: 1(1%) SG2: 1 (1%) Calculated P=1.0

NR

Gerszten (2010)39 Some concerns (6w outcomes) High (12w and later outcomes)

SG1: Plasma disc decompression with coblation technology N randomized: 46 N analyzed: 29 (64% of ITT sample) at 26w NS1: Epidural steroid injection N randomized: 44 N analyzed: 28 (70% of ITT sample) at 26w

All-cause mortality At 26w SG1: 1 (2.2%) NS1: 1 (2.3%) (Causes of death: myocardial infarction, acute pyelonephritis) Surgical mortality SG1: 0 (0%) NS1: 0 (0%)

N (%) with procedure-related adverse events SG1: 5 (11%) NS1: 7 (18%) Calculated P=0.55

NR Reduction in use of narcotics, at 26w Reported as not significantly different between groups

(continued)








SG1: Automated percutaneous discectomy, endoscopic percutaneous discectomy N randomized: 21 N analyzed: 17 (81.0%) at 26w SG2: Discectomy N randomized: 13 N analyzed: 10 (76.9%) at 26w


NR NR NR




N (%) with surgical infection Overall: 5 (6.3%) NR by group

NR NR



All-cause mortality NR Surgical mortality SG1: 0 (0%) SG2: 0 (0%)

N (%) with spinal fluid leakage SG1: 0 (0%) SG2: 1 (3.3%) N (%) with infection SG1: 0 (0%) SG2: 0 (0%) N (%) with neurovascular injuries SG1: 0 (0%) SG2: 0 (0%)

N (%) with reoperations SG1: 1 (3.3%) (for the treatment of mild lateral stenosis that had not been recognized at the time of the index operation) SG2: 2 (6.7%) (1 for repair of dural sac/spinal fluid leak and 1 for persistent radicular symptoms) Calculated P=1.0

Duration of use of narcotics postoperatively, mean in weeks (range) SG1: 1 (0.43 to 2) SG2: 3.65 (1 to 8)

(continued)










N (%) with nerve root sleeve tear SG1: 1 (10%) SG2: 0 (0%) N (%) with superficial wound infection SG1: 0 (0%) SG2: 1 (8.3%)

NR NR

Mayer (1993)32 High



N (%) with complications SG1: 0 (0%) SG2: 0 (0%)

N (%) with reoperations SG1: 3 (15%) (had percutaneous endoscopic discectomy and later microsurgical discectomy, 2 because of lack of improvement and 1 due to recurrence of symptoms) SG2: 1 (5%) (due to epidural scar tissue and progressive neurological symptoms) Calculated P=0.61

NR

(continued)









All-cause mortality NR Surgical mortality SG1: 0 (0%) NS1: NA

N (%) with new neurologic deficits SG1: 0 (0%) NS1: 0 (0%) N (%) with significant adverse events SG1: 0 (0%) NS1: 0 (0%)

N (%) with reoperations SG1: 0 (0%) NS1: NA

NR

North (2005)47 High


All-cause mortality At 26w: SG1: 0 (0%) NS1: 1 (2%) Surgical mortality SG1: 0 (0%) NS1: 0 (0%)

N (%) with site infection SG1: 0 (0%) NS1: 1 (4.2%)

N (%) with reoperations (SG1) or hardware revisions (NS1) SG1: None explicitly reported NS1: 3 (12.5%)

Long-term followup occurred at a mean of 2.9y (range 1.8 to 5.7) N (%) with opioid use stable or decreased at long-term followup SG1: 15 (58%) NS1: 30 (87%) P = 0.025

(continued)










N (%) with urosepsis SG1: 1 (3.6%) No other operative complications were noted

N (%) with reoperations SG1: 2 (7.1%) (due to recurring symptoms on the same side and level) NS1: NA

NR


SG1: Microdiscectomy N randomized: 141 N analyzed: 52w: 140 (99.3%) 2y: 130 (92.2%) 5y: 115 (81.6%) NS1: Conservative management N randomized: 142 N analyzed: 52w: 141 (99.3%) 2y: 130 (91.5%) 5y: 116 (81.7%)


N (%) with surgical complications SG1: 3 (1.6%) (2 dural tears and 1 wound hematoma) NS1: NA

N (%) with reoperation for recurrent sciatica 52w: S1 NR (3.2%); NS 1 NR (1.8% in crossovers who underwent surgery) 2y: SG1 7 (6%); NS1 4 (6% in crossovers who underwent surgery); Calculated P=0.54 5y: SG1 9 (7%); NS1 8 (12% in crossovers who underwent surgery); Calculated P=0.81

NR

(continued)









All-cause mortality Overall: 1 (0.5%) (unrelated to operation) Surgical mortality SG1: 0 (0%) SG2: 0 (0%)

Overall, complication rate reported to be significantly elevated in SG2 as compared to SG1 (P < 0.05). N (%) with dural/nerve injury SG1: 0 (0%) SG2: 0 (0%) N (%) with cauda equina syndrome SG1: 0 (0%) SG2: 0 (0%) N (%) with transient postoperative dysesthesia SG1: 3 (3.3%) SG2: 5 (5.7%) N (%) with postoperative bleeding SG1: 0 (0%) SG2: 2 (2.3%) N (%) with delayed wound healing SG1: 0 (0%) SG2: 1 (1.1%) N (%) with soft tissue infection SG1: 0 (0%) SG2: 1 (1.1%) N (%) with thrombosis SG1: 0 (0%) SG2: 0 (0%) N (%) with spondylodiscitis SG1: 0 (0%) SG2: 0 (0%)

N (%) with revision surgery SG1: 7 (7.7%) SG2: 10 (11.5%) Calculated P=0.45

NR

(continued)









All-cause mortality SG1: 0 (0%) SG2: 0 (0%) Surgical mortality SG1: 0 (0%) SG2: 0 (0%)

N (%) with dural injury SG1: 1 (2%) SG2: 3 (6%) N (%) with nerve injury, cauda-equina syndrome, spondylodiscitis, or thrombosis SG1: 0 (0%) SG2: 0 (0%) N (%) with transient postoperative dysesthesia SG1: 2 (4%) SG2: 5 (10%) N (%) with delayed wound healing SG1: 0 (0%) SG2: 2 (4%) N (%) with soft tissue infection SG1: 0 (0%) SG2: 1 (2%) All morbidity above, calculated P=0.02 Serious complication (not further described) SG1: NR (6%) SG2: NR (21%) P < 0.05

N (%) with revision with conventional spinal canal decompression due to persistent leg pain SG1: 2 (4%) SG2: 1 (2%) Calculated P=1.0 N (%) with revision with fusion due to progradient back pain SG1: 0 (0%) SG2: 2 (4%) Calculated P=0.23 All reoperations, calculated P=0.67

Postoperative pain medication was significantly reduced in SG1 compared with SG2 P < 0.01

(continued)










N (%) with dural tear SG1: 0 (0%) SG2: 2 (6.6%) Calculated P=0.49

N (%) with reoperations for recurrent herniation SG1: 2 (6.6%) (one during initial hospital stay and one at 6w) SG2: 4 (13.3%) (one during initial hospital stay and 1 at 8w, 28w, and 1.2y) Calculated P=0.67

NR




NR NR NR

(continued)









All-cause mortality NR Surgical mortality SG1: 0 (0%) SG2: 0 (0%) SG3: 0 (0%)

N (%) with dural tear SG1: 6 (8.7%) SG2: 2 (2.7%) SG3: 2 (3%) P=0.37 N (%) with root injury SG1: 2 (3%) SG2: 0 (0%) SG3: 0 (0%) P=0.45 N (%) with wound infection SG1: 0 (0%) SG2: 4 (5.5%) SG3: 3 (4.2%) P= 0.29 N (%) with spondylodiscitis SG1: 1 (1.4%) SG2: 0 (0%) SG3: 0 (0%) P=0.56 N (%) with worsening motor deficit SG1: 2 (1.4%) SG2: 1 (1%) SG3: 0 (0%) P=0.47

N (%) with reoperations at 2y Overall: 15 (7%) (13 for recurrences and 2 for repair of pseudomeningocele) N (%) reoperations for recurrence at 2y SG1: 8 (11.4%) SG2: 3 (4.2%) SG2: 2 (3%) Calculated SG1 vs SG2 P=0.13 Calculated SG2 vs SG3 P=1.00

NR

(continued)










N (%) with intraoperative complications SG1: 0 (0%) SG2: 0 (0%) N (%) with superficial wound infection SG1: 0 (0%) SG2: 1 (2.4%)

N (%) with reoperation for reherniation at 1.5y SG1: 2 (5%) SG2: 4 (10%) Calculated P=0.68 [Note: Reherniations (with or without reoperations) were reported in 12.5% of SG1 and 10.5% in SG2 by 2y]

NR




N (%) with nerve root sheath tear SG1: 1 (3.3%) SG2: 1 (3.3%) Calculated P=1.0 N (%) with dural leak SG1: 1 (3.3%) SG2: 0 (0%) Calculated P=1.0 N (%) with discitis SG1: 1 (3.3%) SG2: 0 (0%) Calculated P=1.0 All morbidity, calculated P=0.61

N (%) with reoperation at 52w SG1: 1 (3.3%) at 16w SG2: 1 (3.3%) at <52w Calculated P=1.0

NR

Weber (1983)5 High


All-cause cumulative mortality 52w: NR 4y: SG1 1 (1.7%); NS1 0 (0%) 10y: SG1 3 (5.0%); NS1 0 (0%) Surgical mortality NR

NR NR NR

(continued)







Weinstein (2006)21 Weinstein (2008)45 Lurie Jon (2014)44 High

SG1: Discectomy/ microdiscectomy N randomized: 245 N analyzed: N=232 in main study's primary analyses. 52w: 202 (82.47%) 2y: 186 (75.9%) 3y: 180 (73.5%) 4y: 149 (60.8%) 8y: 157 (64.1%) NS1: Conservative management N randomized: 256 N analyzed: N=240 included in main study's primary analyses. 52w: 213 (83.2%) 2y: 187 (73.0%) 3y: 170 (66.4%) 4y: 150 (58.6%) 8y: 152 (59.4%)

N (%) cumulative all-cause mortality 6w, 12w, 26w: SG1 0 (0%); NS1 0 (0%) 52w: SG1 0 (0%); NS1 1 (0.47%) 2y: SG1 0 (0%); NS1 2 (1.07%) 4y: SG1 1 (0.67%); NS1 2 (1.33%) 8y: SG1 3 (1.91%); NS1 4 (2.63%) N (%) with perioperative death SG1: 0 (0%) NS1: NA

Based on primary analyses in main study for N=247 (SG1 140; NS1 107) participants that had surgery by 2y N (%) with dural tear/spinal fluid leak SG1: 10 (4.0%) N (%) with vascular injury SG1: 1 (0.40%) N (%) with other intraoperative complication SG1: 2 (0.81%) N (%) with postoperative wound infection, superficial SG1: 4 (1.6%) N (%) with other complication (unspecified) SG1: 9 (3.6%)

N (%) with reoperations by 2y 1y: 18 (7.3%) 2y: 25 (10.1%) Reoperations for additional surgery, recurrent herniation, complication, or other reason in the N=247 participants (SG1 140; NS1 107) that had initial surgery by 2y.

NR

Abbreviations: N = number; NR = not reported; NS = not significant; NS1 = nonsurgical intervention group; SG1 = surgical intervention group; w = week(s); y = year(s)



Table C-6. Individual study findings related to cost outcomes

Cost Study Author (Year); Main Study Author (Year); Country

Intervention [SG1] (N randomized); Comparator(s) [SG2, SG3, NS1] (N randomized) Study Methods Results (As Reported by Study) Results (Converted to 2010 U.S. Dollars)a

Malter (1996)42 United States, but used efficacy estimates from non-U.S. trial.

NA Study design: CEA Year/unit of currency reported: 1993 USD Discount rate: 5% Time horizon: 10y Costs included: Direct medical costs (inpatient, outpatient, medication, diagnostic services, other health care services) obtained between 1987 and 1989 from a commercial U.S. database. QOL measure(s) used: Author developed time-tradeoff utility measure Other: Efficacy estimates based on an RCT comparing surgery with nonsurgical treatment (Weber et al (1983)5), and an RCT comparing surgery with chemonucleolysis (Javid et al. (1998)106) and a cohort study comparing surgery with nonsurgical treatment [Atlas et al. (1993)107].

QALY at 10y (undiscounted) SG1: 8.70 (95% CI, NR) NS1: 8.27 (95% CI, NR) AMD: 0.43 (95% CI, NR) Costs at 10y SG1: $17,020 (95% CI, NR) NS1: $4,470 (95% CI, NR) AMD: $12,550 (95% CI, NR) Cost/QALY gained at 10y (undiscounted, payor perspective) $29,200 (95% CI, NR) Cost/QALY gained at 10y (discounted, payor perspective) $33,900 (95% CI, NR)

QALY at 10y (undiscounted) SG1: 8.70 (95% CI, NR) NS1: 8.27 (95% CI, NR) AMD: 0.43 (95% CI, NR) Costs at 10y SG1: $25,684 (95% CI, NR) NS1: $6,745 (95% CI, NR) AMD: $18,938 (95% CI, NR) Cost/QALY gained at 10y (undiscounted, payor perspective) $44,064 (95% CI, NR) Cost/QALY gained at 10y (discounted, payor perspective) $51,156 (95% CI, NR)

(continued)



Table C-6. Individual study findings related to cost outcomes (continued)



Stevenson (1995)53 Chatterjee (1995)36 United Kingdom

Automated percutaneous lumbar discectomy (31); Microdiscectomy (40)

Study design: CEA concurrent to RCT Year/unit of currency reported: 1992 GBP Discount rate: NR Time horizon: 26w Costs included: Direct medical costs of procedures (inpatient, outpatient, medication, capital equipment), patient-reported direct and indirect costs and social service usage (e.g., missed work, travel, paid and unpaid caregivers or domestic help) QOL measure(s) used: NR Other: effectiveness was assessed on a 4-pt Likert scale by two clinicians (4=excellent, 1=poor). "Successful outcome" was defined as a 3 or 4.

Mean total cost at 26w SG1: £2,317 (95% CI NR) SG2: £1,567 (95% CI NR) AMD: NR Calculated AMD: £750 Note: SG1 includes cost of additional microdiscectomy in failed cases and SG2 includes cost of repeat microdiscectomy in failed case. Cost per successful outcome at 26w SG1: £3,264 SG2: £1,958 Cost per point gained on 4-pt Likert scale of Effectiveness SG1: £1,381 SG2: £764

Mean total cost at 26w SG1: $6,340 (95% CI NR) SG2: $4,288 (95% CI NR) AMD: NR Calculated AMD: $2,052 Note: SG1 includes cost of additional microdiscectomy in failed cases and SG2 includes cost of repeat microdiscectomy in failed case. Cost per successful outcome at 26w SG1: $8,931 SG2: $5,358 Calculated AMD: $3,573 Cost per point gained on 4-pt Likert scale of effectiveness SG1: $3,779 SG2: $2,091 Calculated AMD: $1,688 (95% CI cannot be calculated)

Teli (2010)27 (main study includes cost) Italy

Microendoscopic discectomy (70) Microdiscectomy (72) Open discectomy (70)

Study design: Cost analysis concurrent to RCT Year/unit of currency reported: Euros, Year NRb Discount rate: NR Time horizon: NA Costs included: direct surgical equipment costs (equipment, tools), operating times, rehospitalizations QOL measure(s) used: NA

Mean surgical costs (SD) SG1: €3,010 (450) SG2: € 2,450 (340) SG3: €2,310 (260) Two P values were provided (P=0.002 and P=0.012), but unclear what comparison they are referring to.

Mean surgical costs (SD) SG1: $3,878 ($580) SG2: $3,156 ($438) SG3: $2,976 ($322) Calculated AMD SG1 vs. SG2: $722 (95% CI, $551 to $892) Calculated AMD SG2 vs. SG3: $65 (95% CI $52 to $307) Calculated AMD SG1 vs. SG3: $902 (95% CI $745 to $1059

(continued)






Tosteson (2008)50 Weinstein (2006)21 SPORT United States

Discectomy (245); Conservative management (256)

Study design: CEA concurrent to RCT Year/unit of currency: 2004 USD Discount rate: 3% Time horizon: 2y Costs included: Direct medical costs (inpatient, outpatient, medication, diagnostic services, other health care services), patient-reported indirect costs (e.g., missed work, unpaid caregivers) QOL measure(s) used: EQ-5D with US scoring Other: Based on pooled data from SPORT RCT and observational cohort. Crossovers considered in the 'as treated' group. Total N 1,191 (775 surgery, 416 no surgery)

Mean discounted QALYs at 2y SG1: 1.64 (95% CI, 1.62 to 1.67) NS1: 1.44 (95% CI, 1.41. to 1.47) AMD: 0.21 (95% CI, 0.16 to 25) Mean total costs at 2y SG1: $27,341 (95% CI, $25,882 to $28,799) NS1: $13,135 (95% CI, 11,244 to $14,902) Calculated AMD: $14,206 Cost/QALY gained (societal perspective) at 2y $69,403 (95% CI, $49,523 to $94,999) Direct medical costs at 2y SG1: $20,237 ($19,314 to $21,160) NS1: $5,804 (95% CI, $4,639 to $6,969) Calculated AMD: $14,433 Direct medical costs/QALY gained (payor perspective) at 2y $72,181 (95% CI, $56,473 to $92,394)

Mean discounted QALYs at 2y SG1: 1.64 (95% CI, 1.62 to 1.67) NS1: 1.44 (95% CI, 1.41. to 1.47) AMD: 0.21 (95% CI, 0.16 to 25) Mean total costs at 2y SG1: $31,561 (95% CI, $29,877 to $33,244) NS1: $15,162 (95% CI, $12,979 to $17,202) Calculated AMD: $16,399 (95% CI, $16,289 to $16,509) Cost/QALY gained (societal perspective) at 2y $80,115 (95% CI, $57,167 to $109,662) Direct medical costs at 2y SG1: $23,361 ($22,295 to $24,426) NS1: $6,700 (95% CI, $5,355 to $8,045) Calculated AMD: $16,661 (95% CI, $16,590 to $16,732) Direct medical costs/QALY gained (payor perspective) at 2y $83,322 (95% CI, $65,189 to $106,655)

(continued)






Van den Akker (2011)51 Arts (2009)38 The Sciatica Micro-Endoscopic Discectomy Randomized Controlled Trial The Netherlands

Tubular discectomy (167); Microdiscectomy (161)

Study design: CEA concurrent with RCT Year/unit of currency reported: 2008 USD Discount rate: 0% Time horizon: 52w Costs included: Direct medical costs (inpatient, outpatient, medication, diagnostic services, other health care services), patient-reported indirect costs (e.g., missed work, travel, paid and unpaid caregivers or domestic help) QOL measure(s) used: EQ-5D

QALYs at 52w SG1: NR SG2: NR AMD: -0.012 (95% CI, -0.046 to 0.021) Mean total costs at 52w SG1: $16,858 (SD $12,759) SG2: $15,637 (SD $12,165) AMD: $1,491 (95% CI, $-1,335 to $4,318) Cost/QALY gained (societal perspective) at 52w NR. Differences in costs and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest microdiscectomy dominates minimally invasive surgery (i.e., is more effective and costs less). Mean health care costs at 52w SG1: $5,529 (SD $3,020) SG2: $5,070 ($3,375) AMD: $460 (95% CI, $-243 to $1,163) Health care costs/QALY gained (payor perspective) at 52w NR. Differences in costs and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest microdiscectomy dominates minimally invasive surgery (i.e., is more effective and costs less).

QALYs at 52w SG1: NR SG2: NR AMD: -0.012 (95% CI, -0.046 to 0.021) Mean total costs at 52w SG1: $17,074 (SD $12,922) SG2: $15,837 (SD $12,321) AMD: $1,510 (95% CI, -$1,352 to $4,373) Cost/QALY gained (societal perspective) at 52w NR. Differences in costs and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest microdiscectomy dominates minimally invasive surgery (i.e., is more effective and costs less). Calculated cost per QALY gained $-125,833 (95% CI cannot be calculated) Mean health care costs at 52w SG1: $5,600 (SD $3,059) SG2: $5,135 (SD $3,418) AMD: $466 (95% CI, $-246 to $1,178) Health Care Costs/QALY Gained (payor perspective) at 52w NR. Differences in costs and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest microdiscectomy dominates minimally invasive surgery (i.e., is more effective and costs less). Calculated cost per QALY gained $-38,833 (95% CI cannot be calculated).

(continued)






Van den Akker (2017)52 Brouwer (2015)37 The Netherlands

Percutaneous laser disc decompression (57); Microdiscectomy (58)

Study design: CEA concurrent to RCT Year/unit of currency reported: 2010 Euros Discount rate: 0% Time horizon: 52w Costs included: Direct medical costs (inpatient, outpatient, medication, diagnostic services, other health care services), patient-reported indirect costs (e.g., missed work, travel, paid and unpaid caregivers or domestic help) QOL measure(s) used: EQ-5D with US Scoring

QALY at 52w SG1: 0.733 (SD 0.172) SG2: 0.766 (0.133) AMD: -0.033 (95% CI, NR, P=0.27) Mean total costs at 52w SG1: €18,071 (SD €14,351) SG2: €20,451 (SD €13,080) AMD: €-2,379 (95% CI, €-7,618 to €2,860) Costs/QALY gained (societal perspective) at 52w NR. Differences in cost and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest minimally invasive surgery may be less effective but also costs less. Mean health care costs at 52w SG1: €5,325 (SD €4,395) SG2: €7,095 (SD €3,109) AMD: €-1,771 (95% CI, €-3,238 to €-303) Health care costs/QALY gained (payor perspective) at 52w NR. Difference in QALYs underpinning this calculated value was not statistically significant, but point estimates suggest minimally invasive surgery may be less effective but also costs less.

QALY at 52w SG1: 0.733 (SD 0.172) SG2: 0.766 (SD 0.133) AMD: -0.033 (95% CI, NR, P=0.27) Mean total costs at 52w SG1: $24,420 (SD $19,393) SG2: $27,636 (SD $17,676) AMD: $-3,215 (95% CI, $-10,294 to $3,865) Costs/QALY gained (societal perspective) at 52w NR. Differences in cost and QALYs underpinning this calculated value were not statistically significant, but point estimates suggest minimally invasive surgery may be less effective but also costs less. Calculated cost per QALY gained for microdiscectomy compared to minimally invasive surgery $97,424 (95% CI cannot be calculated) Mean health care costs at 52w SG1: $7,196 (SD $5,939) SG2: $9,588 (SD $4,201) AMD: $-2,393 (95% CI, $-4,376 to $-409) Health care costs/QALY gained (payor perspective) at 52w NR. Difference in QALYs underpinning this calculated value was not statistically significant, but point estimates suggest minimally invasive surgery may be less effective but also costs less. Calculated health care costs cost per QALY gained for microdiscectomy compared to minimally invasive surgery $72,515 95% CI cannot be calculated).

(continued)






Van den Hout (2008)49 Peul (2007)30 Sciatica Trial The Netherlands

Discectomy (141); Conservative management (142)

Study design: CEA concurrent to RCT Year/unit of currency reported: 2008 Euros Discount rate: 0% Time horizon: 52w Costs included: Direct medical costs (inpatient, outpatient, medication, diagnostic services, other health care services), patient-reported indirect costs (e.g., missed work, travel, paid and unpaid caregivers or domestic help) QOL measure(s) used: EQ-5D with UK scoring

Mean QALY at 52w SG1: 0.78 (SD 0.17) NS1: 0.73 (SD 0.16) AMD: 0.044 (95% CI, 0.005 to 0.083) Mean total costs at 52w SG1: €18,493 (SD €14,548) NS1: €18,506 (SD €18,102) AMD: €-12 (95% CI, €-4,029 to €4,006) Cost/QALY gained (societal perspective) at 52w NR, but SG1 dominates NS1 (more effective and less costs) but differences in cost underpinning this calculated value were not statistically significant. Calculated to be €-272. Using lower and upper 95% CI on AMDs for QALY and cost, best case ICER is calculated to be €-48,542 and worst case ICER is calculated to be €801,200 Mean health care costs at 52w SG1: €5,626 (SD €3,875) NS1: €3,807 (SD€4,237) AMD: €1,819 (95% CI, €842 to €2,790) Health care costs/QALY gained (payor perspective) at 52w €41,000 (95% CI, €14,000 to €430,000)

Mean QALY at 52w SG1: 0.78 (SD 0.17) NS1: 0.73 (SD 0.16) AMD: 0.044 (95% CI, 0.005 to 0.083) Mean total costs at 52w SG1: $28,421 (SD $22,358) NS1: $28,441 (SD $27,820) AMD: $-18.44 (95% CI, $-6,192 to $6,157) Cost/QALY gained (societal perspective) at 52w NR, but SG1 dominates NS1 (more effective and less costs) but differences in cost underpinning this calculated value were not statistically significant. Calculated to be $-419. Using lower and upper 95% CI on AMDs for QALY and cost, best case ICER is calculated to be $-74,602 and worst case ICER is calculated to be $1,231,400. Mean health care costs at 52w SG1: $8,646 (SD $5,955) NS1: $5,851 (SD $6,512) AMD: $2,796 (95% CI, $1,294 to $4,288) Health care costs/QALY gained (payor perspective) at 52w $63,011 (95% CI, $21,516 to $660,847)

a See Appendix B for description of methods used to convert costs to 2010 U.S. dollars.

b The study did not provide the year; we assumed 2009 for purposes of converting to 2010 U.S. dollars.

Abbreviations: AMD = absolute mean difference; CEA = cost-effectiveness analysis; CI = confidence interval; NA = not applicable; ICER = incremental cost-effectiveness ratio;

NR = not reported; NS = nonsurgical group; QALY = quality-adjusted life year; w = weeks(s); QOL = quality of life; RCT = randomized controlled trial; SD = standard deviation;

SG = surgical group; USD = United States Dollar; y = year


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page D-1

Appendix D. Excluded Studies

List of Exclusion Codes

X1: Systematic review for handsearch

X2: Ineligible publication type

X3: Ineligible country

X4: Ineligible population

X5: Ineligible intervention

X6: Ineligible comparator

X7: Ineligible outcome

X8: Ineligible study design

X9: Duplicate or superseded

X10: Study protocol or in progress

X11: Abstract only

X12: Non-English full text

X13: Data uninterpretable

1. Anulex Technologies I. Randomized Study of Anular

Repair With the Xclose Tissue Repair System. 2008.

Exclusion Code: X10.

2. University of Sao Paulo. Percutaneous Diskectomy

SpineJet x Open Microdiskectomy in Treatment of

Lumbar Radiculopathy (PDOP_TLR). 2011.


3. Abdu RW, Abdu WA, Pearson AM, et al.

Reoperation for Recurrent Intervertebral Disc

Herniation in the Spine Patient Outcomes Research

Trial: Analysis of Rate, Risk Factors, and Outcome.

Spine (Phila Pa 1976). 2017 Jul 15;42(14):1106-14.

doi: 10.1097/brs.0000000000002088. PMID:

28146015. Exclusion Code: X8.

4. Abramovitz JN, Neff SR. Lumbar disc surgery:

results of the Prospective Lumbar Discectomy Study

of the Joint Section on Disorders of the Spine and

Peripheral Nerves of the American Association of

Neurological Surgeons and the Congress of

Neurological Surgeons. Neurosurgery. 1991

Aug;29(2):301-7; discussion 7-8. PMID: 1886676.

Exclusion Code: X8.

5. Adakli B, Cakar Turhan KS, Asik I. The comparison

of the efficacy of radiofrequency nucleoplasty and

targeted disc decompression in lumbar radiculopathy.

Bosn J Basic Med Sci. 2015 Apr 25;15(2):57-61. doi:

10.17305/bjbms.2015.427. PMID: 26042514.

Exclusion Code: X8.

6. Adam D, Pevzner E, Gepstein R. Comparison of

percutaneous nucleoplasty and open discectomy in

patients with lumbar disc protrusions. Chirurgia

(Bucur). 2013 Jan-Feb;108(1):94-8. PMID:


7. Ahn SS, Kim SH, Kim DW, et al. Comparison of

Outcomes of Percutaneous Endoscopic Lumbar

Discectomy and Open Lumbar Microdiscectomy for

Young Adults: A Retrospective Matched Cohort

Study. World Neurosurg. 2016 Feb;86:250-8. doi:

10.1016/j.wneu.2015.09.047. PMID: 26409086.

Exclusion Code: X8.

8. Amoretti N, Huwart L, Marcy PY, et al. CT- and

fluoroscopy-guided percutaneous discectomy for

lumbar radiculopathy related to disc herniation: a

comparative prospective study comparing lateral to

medial herniated discs. Skeletal Radiol. 2013

Jan;42(1):49-53. doi: 10.1007/s00256-012-1422-5.

PMID: 22644540. Exclusion Code: X8.

9. Anderson DG, Patel A, Maltenfort M, et al. Lumbar

decompression using a traditional midline approach

versus a tubular retractor system: comparison of

patient-based clinical outcomes. Spine (Phila Pa

1976). 2011 Mar 01;36(5):E320-5. doi:

10.1097/BRS.0b013e3181db1dfb. PMID: 21178844.

Exclusion Code: X8.

10. Anderson PA, McCormick PC, Angevine PD.

Randomized controlled trials of the treatment of

lumbar disk herniation: 1983-2007. J Am Acad

Orthop Surg. 2008 Oct;16(10):566-73. PMID:


11. Anichini G, Landi A, Caporlingua F, et al. Lumbar

endoscopic microdiscectomy: where are we now? An

updated literature review focused on clinical

outcome, complications, and rate of recurrence.

Biomed Res Int. 2015;2015:417801. doi:

10.1155/2015/417801. PMID: 26688809. Exclusion

Code: X2.

12. Arai Y, Hirai T, Yoshii T, et al. A prospective

comparative study of 2 minimally invasive

decompression procedures for lumbar spinal canal

stenosis: unilateral laminotomy for bilateral

decompression (ULBD) versus muscle-preserving

interlaminar decompression (MILD). Spine (Phila Pa

1976). 2014 Feb 15;39(4):332-40. doi:

10.1097/brs.0000000000000136. PMID: 24299721.

Exclusion Code: X4.



13. Aronsohn J, Chapman K, Soliman M, et al.

Percutaneous microdiscectomy versus epidural

injection for management of chronic spinal pain. Proc

West Pharmacol Soc. 2010;53:16-9. PMID:


14. Arts M, Brand R, van der Kallen B, et al. Does

minimally invasive lumbar disc surgery result in less

muscle injury than conventional surgery? A

randomized controlled trial. Eur Spine J. 2011

Jan;20(1):51-7. doi: 10.1007/s00586-010-1482-y.


15. Arts MP, Brand R, Koes BW, et al. Effect modifiers

of outcome of surgery in patients with herniated disc

related sciatica? A subgroup analysis of a randomised

clinical trial. J Neurol Neurosurg Psychiatry. 2010

Nov;81(11):1265-74. doi:

10.1136/jnnp.2009.192906. Epub 2010 Jun 14.


16. Arts MP, Peul WC, Brand R, et al. Cost-effectiveness

of microendoscopic discectomy versus conventional

open discectomy in the treatment of lumbar disc

herniation: a prospective randomised controlled trial

[ISRCTN51857546]. BMC Musculoskelet Disord.

2006 May 13;7:42. doi: 10.1186/1471-2474-7-42.


17. Atlas SJ, Tosteson TD, Blood EA, et al. The impact

of workers' compensation on outcomes of surgical

and nonoperative therapy for patients with a lumbar

disc herniation: SPORT. Spine (Phila Pa 1976). 2010

Jan 01;35(1):89-97. doi:

10.1097/BRS.0b013e3181c68047. PMID: 20023603.

Exclusion Code: X8.

18. Ay S, Dogan SK, Evcik D. Is low-level laser therapy

effective in acute or chronic low back pain? Clin

Rheumatol. 2010 Aug;29(8):905-10. doi:

10.1007/s10067-010-1460-0. PMID: 20414695.

Exclusion Code: X5.

19. Aykac B, Copuroglu C, Ozcan M, et al. Postoperative

evaluation of quality of life in lumbar spinal stenosis

patients following instrumented posterior

decompression. Acta Orthop Traumatol Turc.

2011;45(1):47-52. doi: 10.3944/aott.2011.2387.


20. Azarhomayoun A, Chou R, Shirdel S, et al.

Sequestrectomy Versus Conventional

Microdiscectomy for the Treatment of a Lumbar Disc

Herniation: A Systematic Review. Spine (Phila Pa

1976). 2015 Dec;40(24):E1330-9. doi:

10.1097/brs.0000000000001174. PMID: 26655808.

Exclusion Code: X1.

21. Azzazi A, AlMekawi S, Zein M. Lumbar disc

nucleoplasty using coblation technology: clinical

outcome. J Neurointerv Surg. 2011 Sep;3(3):288-92.

doi: 10.1136/jnis.2010.002402. PMID: 21990844.

Exclusion Code: X8.

22. Bailey A, Araghi A, Blumenthal S, et al. Prospective,

multicenter, randomized, controlled study of anular

repair in lumbar discectomy: two-year follow-up.

Spine (Phila Pa 1976). 2013 Jun 15;38(14):1161-9.

doi: 10.1097/BRS.0b013e31828b2e2f. PMID:


23. Banczerowski P, Czigleczki G, Papp Z, et al.

Minimally invasive spine surgery: systematic review.

Neurosurg Rev. 2015 Jan;38(1):11-26; discussion

doi: 10.1007/s10143-014-0565-3. PMID: 25199809.

Exclusion Code: X2.

24. Barth M, Diepers M, Weiss C, et al. Two-year

outcome after lumbar microdiscectomy versus

microscopic sequestrectomy: part 2: radiographic

evaluation and correlation with clinical outcome.

Spine (Phila Pa 1976). 2008 Feb 01;33(3):273-9. doi:

10.1097/BRS.0b013e31816201a6. PMID: 18303459.

Exclusion Code: X7.

25. Bokov A, Skorodumov A, Isrelov A, et al.

Differential treatment of nerve root compression pain

caused by lumbar disc herniation applying

nucleoplasty. Pain Physician. 2010 Sep-

Oct;13(5):469-80. PMID: 20859316. Exclusion

Code: X8.

26. Boult M, Fraser RD, Jones N, et al. Percutaneous

endoscopic laser discectomy. Aust N Z J Surg. 2000

Jul;70(7):475-9. PMID: 10901572. Exclusion Code:

X8.

27. Bron JL, Helder MN, Meisel HJ, et al. Repair,

regenerative and supportive therapies of the annulus

fibrosus: achievements and challenges. Eur Spine J.

2009 Mar;18(3):301-13. doi: 10.1007/s00586-008-

0856-x. PMID: 19104850. Exclusion Code: X2.

28. Brouwer PA, Peul WC, Brand R, et al. Effectiveness

of percutaneous laser disc decompression versus

conventional open discectomy in the treatment of

lumbar disc herniation; design of a prospective

randomized controlled trial. BMC Musculoskelet

Disord. 2009 May 13;10:49. doi: 10.1186/1471-

2474-10-49. PMID: 19439098. Exclusion Code: X10.

29. Buttermann GR. Treatment of lumbar disc herniation:

epidural steroid injection compared with discectomy.

A prospective, randomized study. J Bone Joint Surg

Am. 2004 Apr;86-a(4):670-9. PMID: 15069129.

Exclusion Code: X8.

30. Chang X, Chen B, Li HY, et al. The safety and

efficacy of minimally invasive discectomy: a meta-

analysis of prospective randomised controlled trials.

Int Orthop. 2014 Jun;38(6):1225-34. doi:

10.1007/s00264-014-2331-0. PMID: 24722785.

Exclusion Code: X1.

31. Chitragran R, Poopitaya S, Tassanawipas W. Result

of percutaneous disc decompression using

nucleoplasty in Thailand: a randomized controlled

trial. J Med Assoc Thai. 2012 Oct;95 Suppl 10:S198-

205. PMID: 23451463. Exclusion Code: X3.



32. Choi G, Lee SH, Deshpande K, et al. Working

channel endoscope in lumbar spine surgery. J

Neurosurg Sci. 2014 Jun;58(2):77-85. PMID:


33. Chou R, Baisden J, Carragee EJ, et al. Surgery for

low back pain: a review of the evidence for an

American Pain Society Clinical Practice Guideline.

Spine (Phila Pa 1976). 2009 May 01;34(10):1094-

109. doi: 10.1097/BRS.0b013e3181a105fc. PMID:


34. Chou R, Loeser JD, Owens DK, et al. Interventional

therapies, surgery, and interdisciplinary rehabilitation

for low back pain: an evidence-based clinical practice

guideline from the American Pain Society. Spine

(Phila Pa 1976). 2009 May 01;34(10):1066-77. doi:

10.1097/BRS.0b013e3181a1390d. PMID: 19363457.

Exclusion Code: X2.

35. Clark AJ, Safaee MM, Khan NR, et al. Tubular

microdiscectomy: techniques, complication

avoidance, and review of the literature. Neurosurg

Focus. 2017 Aug;43(2):E7. doi:

10.3171/2017.5.focus17202. PMID: 28760036.

Exclusion Code: X1.

36. Cole JSt, Jackson TR. Minimally invasive lumbar

discectomy in obese patients. Neurosurgery. 2007

Sep;61(3):539-44; discussion 44. doi:

10.1227/01.neu.0000290900.23190.c9. PMID:


37. Cong L, Zhu Y, Tu G. A meta-analysis of endoscopic

discectomy versus open discectomy for symptomatic

lumbar disk herniation. Eur Spine J. 2016

Jan;25(1):134-43. doi: 10.1007/s00586-015-3776-6.


38. Couto JM, Castilho EA, Menezes PR.

Chemonucleolysis in lumbar disc herniation: a meta-

analysis. Clinics (Sao Paulo). 2007 Apr;62(2):175-


39. Crawshaw C, Frazer AM, Merriam WF, et al. A

comparison of surgery and chemonucleolysis in the

treatment of sciatica A prospective randomized trial.

Spine (Phila Pa 1976). 1984 Mar;9(2):195-8. PMID:


40. Cristante AF, Rocha ID, MartusMarcon R, et al.

Randomized clinical trial comparing lumbar

percutaneous hydrodiscectomy with lumbar open

microdiscectomy for the treatment of lumbar disc

protrusions and herniations. Clinics (Sao Paulo).

2016 May;71(5):276-80. doi:

10.6061/clinics/2016(05)06. PMID: 27276397.

Exclusion Code: X3.

41. Dasenbrock HH, Juraschek SP, Schultz LR, et al. The

efficacy of minimally invasive discectomy compared

with open discectomy: a meta-analysis of prospective

randomized controlled trials. J Neurosurg Spine.

2012 May;16(5):452-62. doi:

10.3171/2012.1.spine11404. PMID: 22404142.

Exclusion Code: X1.

42. Duarte R, Costa J. Percutaneous laser disc

decompression for lumbar discogenic radicular pain.

Radiología (English Edition). 2012;54(4):336-41.

Exclusion Code: X8.

43. Eichen PM, Achilles N, Konig V, et al. Nucleoplasty,

a minimally invasive procedure for disc

decompression: a systematic review and meta-

analysis of published clinical studies. Pain Physician.

2014 Mar-Apr;17(2):E149-73. PMID: 24658486.

Exclusion Code: X1.

44. Ejeskar A, Nachemson A, Herberts P, et al. Surgery

versus chemonucleolysis for herniated lumbar discs.

A prospective study with random assignment. Clin

Orthop Relat Res. 1983 Apr(174):236-42. PMID:


45. Epstein NE. Should anyone perform percutaneous

endoscopic laser diskectomy and percutaneous

lumbar disc decompressions? Surg Neurol Int.

2016;7(Suppl 42):S1080-s4. doi: 10.4103/2152-

7806.196764. PMID: 28144489. Exclusion Code:

X1.

46. National Institute for Clinical Excellence (NICE).

Interventional Procedure Overview of Endoscopic

Laser Foraminoplasty. London, United Kingdom:

National Institute for Clinical Excellence; 2003. p. IP

Guidance Number IPG031. Exclusion Code: X4.

47. National Institute for Clinical Excellence (NICE).

Interventional Procedure Overview of Laser Lumbar

Discectomy. London, United Kingdom: National

Institute for Clinical Excellence; 2003. p. IP

Guidance Number IPG027. Exclusion Code: X1.

48. Fakouri B, Patel V, Bayley E, et al. Lumbar

microdiscectomy versus sequesterectomy/free

fragmentectomy: a long-term (> 2 y) retrospective

study of the clinical outcome. Clinical Spine Surgery.

2011;24(1):6-10. Exclusion Code: X8.

49. Fakouri B, Shetty NR, White TC. Is sequestrectomy a

viable alternative to microdiscectomy? A systematic

review of the literature. Clin Orthop Relat Res. 2015

Jun;473(6):1957-62. doi: 10.1007/s11999-014-3904-


50. Findlay GF, Hall BI, Musa BS, et al. A 10-year

follow-up of the outcome of lumbar

microdiscectomy. Spine (Phila Pa 1976). 1998 May

15;23(10):1168-71. PMID: 9615370. Exclusion

Code: X8.

51. Fisher C, Noonan V, Bishop P, et al. Outcome

evaluation of the operative management of lumbar

disc herniation causing sciatica J Neurosurg. 2004

Apr;100(4 Suppl Spine):317-24. PMID: 15070138.

Exclusion Code: X8.

52. Fitzsimmons D, Phillips CJ, Bennett H, et al. Cost-

effectiveness of different strategies to manage

patients with sciatica Pain. 2014 Jul;155(7):1318-27.



doi: 10.1016/j.pain.2014.04.008. PMID: 24726924.

Exclusion Code: X9.

53. Fourney DR, Dettori JR, Norvell DC, et al. Does

minimal access tubular assisted spine surgery

increase or decrease complications in spinal

decompression or fusion? Spine (Phila Pa 1976).

2010 Apr 20;35(9 Suppl):S57-65. doi:

10.1097/BRS.0b013e3181d82bb8. PMID: 20407352.

Exclusion Code: X1.

54. Fukui S, Iwashita N, Nitta K, et al. The results of

percutaneous intradiscal high-pressure injection of

saline in patients with extruded lumbar herniated

disc: comparison with microendoscopic discectomy.

Pain Med. 2012 Jun;13(6):762-8. doi:

10.1111/j.1526-4637.2012.01400.x. Epub 2012 May


55. Garg B, Nagraja UB, Jayaswal A Microendoscopic

versus open discectomy for lumbar disc herniation: a

prospective randomised study. J Orthop Surg (Hong

Kong). 2011 Apr;19(1):30-4. doi:

10.1177/230949901101900107. PMID: 21519072.

Exclusion Code: X3.

56. Gerges FJ, Lipsitz SR, Nedeljkovic SS. A systematic

review on the effectiveness of the Nucleoplasty

procedure for discogenic pain. Pain Physician. 2010

Mar-Apr;13(2):117-32. PMID: 20309378. Exclusion

Code: X1.

57. Gibson JA, Waddell G. Surgical interventions for

lumbar disc prolapse. Cochrane Database of

Systematic Reviews: John Wiley & Sons, Ltd; 2007.

Exclusion Code: X9.

58. Gibson JN, Waddell G. Surgical interventions for

lumbar disc prolapse: updated Cochrane Review.

Spine (Phila Pa 1976). 2007 Jul 15;32(16):1735-47.

doi: 10.1097/BRS.0b013e3180bc2431. PMID:


59. Gibson JN, Waddell G. Surgical interventions for

lumbar disc prolapse. Cochrane Database Syst Rev.

2007 Apr 18(2):Cd001350. doi:

10.1002/14651858.CD001350.pub4. PMID:


60. Gotfryd A, Avanzi O. A systematic review of

randomised clinical trials using posterior discectomy

to treat lumbar disc herniations. Int Orthop. 2009

Feb;33(1):11-7. doi: 10.1007/s00264-008-0559-2.


61. . Microdiscectomy and conservative treatment for

lumbar disc herniation with back pain and sciatica: a

randomised clinical trial. Proceedings of the 30th

meeting of the International Society for the Study of

the Lumbar Spine; 2003; Vancouver. Exclusion

Code: X11.

62. Guyatt GH, Cook DJ, Jaeschke R, et al. Grades of

recommendation for antithrombotic agents: American

College of Chest Physicians evidence-based clinical

practice guidelines. CHEST Journal.

2008;133(6_suppl):123S-31S. Exclusion Code: X2.

63. Helm Ii S, Deer TR, Manchikanti L, et al.

Effectiveness of thermal annular procedures in

treating discogenic low back pain. Pain Physician.

2012 May-Jun;15(3):E279-304. PMID: 22622914.

Exclusion Code: X4.

64. Hirsch JA, Singh V, Falco FJ, et al. Automated

percutaneous lumbar discectomy for the contained

herniated lumbar disc: a systematic assessment of

evidence. Pain Physician. 2009 May-Jun;12(3):601-


65. Hussein M. Minimal Incision, Multifidus-sparing

Microendoscopic Diskectomy Versus Conventional

Microdiskectomy for Highly Migrated Intracanal

Lumbar Disk Herniations. J Am Acad Orthop Surg.

2016 Nov;24(11):805-13. doi: 10.5435/jaaos-d-15-


66. Hussein M, Abdeldayem A, Mattar MM. Surgical

technique and effectiveness of microendoscopic

discectomy for large uncontained lumbar disc

herniations: a prospective, randomized, controlled

study with 8 years of follow-up. Eur Spine J. 2014

Sep;23(9):1992-9. doi: 10.1007/s00586-014-3296-9.


67. Iwamoto J, Sato Y, Takeda T, et al. The return to

sports activity after conservative or surgical treatment

in athletes with lumbar disc herniation. Am J Phys

Med Rehabil. 2010 Dec;89(12):1030-5. doi:

10.1097/PHM.0b013e3181f71044. PMID: 20881585.

Exclusion Code: X4.

68. Jacobs WC, Arts MP, van Tulder MW, et al. Surgical

techniques for sciatica due to herniated disc, a

systematic review. Eur Spine J. 2012

Nov;21(11):2232-51. doi: 10.1007/s00586-012-2422-


69. Jacobs WC, van Tulder M, Arts M, et al. Surgery

versus conservative management of sciatica due to a

lumbar herniated disc: a systematic review. Eur Spine

J. 2011 Apr;20(4):513-22. doi: 10.1007/s00586-010-

1603-7. PMID: 20949289. Exclusion Code: X1.

70. Johans SJ, Amin BY, Mummaneni PV. Minimally

invasive lumbar decompression for lumbar stenosis:

review of clinical outcomes and cost effectiveness. J

Neurosurg Sci. 2015 Mar;59(1):37-45. PMID:


71. Jordan J, Konstantinou K, O'Dowd J. Herniated

lumbar disc. BMJ Clin Evid. 2009 Mar 26;2009


72. Jun-liang LIU W-xZ, Guo-yong GAO et al. A

prospective and controlled study of percutaneous

transforaminal endoscopic discectomy versus

fenestration discectomy for lumbar disc herniation.

Chin J Bone Joint. 2014;3:245–50. Exclusion Code:

X12.



73. Kamper SJ, Ostelo RW, Rubinstein SM, et al.

Minimally invasive surgery for lumbar disc

herniation: a systematic review and meta-analysis.

Eur Spine J. 2014 May;23(5):1021-43. doi:

10.1007/s00586-013-3161-2. PMID: 24442183.

Exclusion Code: X1.

74. Kelly RE, Dinner MH, Lavyne MH, et al. The effect

of lumbar disc surgery on postoperative pulmonary

function and temperature. A comparison study of

microsurgical lumbar discectomy with standard

lumbar discectomy. Spine (Phila Pa 1976). 1993

Feb;18(2):287-90. PMID: 8441946. Exclusion Code:

X7.

75. Kerr D, Zhao W, Lurie JD. What Are Long-term

Predictors of Outcomes for Lumbar Disc Herniation?

A Randomized and Observational Study. Clin Orthop

Relat Res. 2015 Jun;473(6):1920-30. doi:

10.1007/s11999-014-3803-7. PMID: 25057116.

Exclusion Code: X9.

76. Koerner JD, Glaser J, Radcliff K. Which Variables

Are Associated With Patient-reported Outcomes

After Discectomy? Review of SPORT Disc

Herniation Studies. Clin Orthop Relat Res. 2015

Jun;473(6):2000-6. doi: 10.1007/s11999-014-3671-1

10.1007/s11999-014-3671-1. PMID: 24818737.

Exclusion Code: X2.

77. Krappel F, Schmitz R, Bauer E, et al. Open or

endoscopic nucleotomy? Results of a prospective,

controlled clinical trial with independent follow up,

MRI and special reference to cost effectiveness.

Offene oder endoskopische nukleotomie—Ergebnisse

einer kontrollierten klinischen studie mit

unabhangiger nachuntersuchung, MRT und unter

besonderer berucksichtigung der kosten-nutzen-

relation. Orthop Prax. 2001;37:164-9. Exclusion

Code: X12.

78. Kreiner DS, Hwang SW, Easa JE, et al. An evidence-

based clinical guideline for the diagnosis and

treatment of lumbar disc herniation with

radiculopathy. Spine J. 2014 Jan;14(1):180-91. doi:

10.1016/j.spinee.2013.08.003. PMID: 24239490.

Exclusion Code: X2.

79. Krugluger J, Knahr K. Chemonucleolysis and

automated percutaneous discectomy--a prospective

randomized comparison. Int Orthop. 2000;24(3):167-


80. Lagarrigue J, Chaynes P. [Comparative study of disk

surgery with or without microscopy. A prospective

study of 80 cases]. Neurochirurgie. 1994;40(2):116-


81. Lavignolle B, Vital JM, Baulny D, et al.

[Comparative study of surgery and chemonucleolysis

in the treatment of sciatica caused by a herniated

disk]. Acta Orthop Belg. 1987;53(2):244-9. PMID:


82. Lee DY, Lee SH. Carbon dioxide (CO2) laser-

assisted microdiscectomy for extraforaminal lumbar

disc herniation at the L5-S1 level. Photomed Laser

Surg. 2011 Aug;29(8):531-5. doi:

10.1089/pho.2010.2954. PMID: 21309702. Exclusion

Code: X6.

83. Lee SH, Chung SE, Ahn Y, et al. Comparative

radiologic evaluation of percutaneous endoscopic

lumbar discectomy and open microdiscectomy: a

matched cohort analysis. Mt Sinai J Med. 2006

Sep;73(5):795-801. PMID: 17008941. Exclusion

Code: X8.

84. Leven D, Passias PG, Errico TJ, et al. Risk Factors

for Reoperation in Patients Treated Surgically for

Intervertebral Disc Herniation: A Subanalysis of

Eight-Year SPORT Data J Bone Joint Surg Am. 2015

Aug 19;97(16):1316-25. doi: 10.2106/jbjs.n.01287.


85. Lewis R, Williams N, Matar HE, et al. The clinical

effectiveness and cost-effectiveness of management

strategies for sciatica: systematic review and

economic model. Health Technol Assess. 2011

Nov;15(39):1-578. doi: 10.3310/hta15390. PMID:


86. Lewis RA, Williams NH, Sutton AJ, et al.

Comparative clinical effectiveness of management

strategies for sciatica: systematic review and network

meta-analyses. Spine J. 2015 Jun 01;15(6):1461-77.

doi: 10.1016/j.spinee.2013.08.049. PMID: 24412033.

Exclusion Code: X1.

87. Li T, Han D, Liu B, et al. Clinical assessment of

reformed lumbar microdiscectomy. Eur J Orthop

Surg Traumatol. 2014 Jan;24(1):23-7. doi:

10.1007/s00590-012-1123-8. PMID: 23412248.

Exclusion Code: X3.

88. Li X, Han Y, Di Z, et al. Percutaneous endoscopic

lumbar discectomy for lumbar disc herniation. J Clin

Neurosci. 2016 Nov;33:19-27. doi:

10.1016/j.jocn.2016.01.043. PMID: 27475315.

Exclusion Code: X1.

89. Li X, Hu Z, Cui J, et al. Percutaneous endoscopic

lumbar discectomy for recurrent lumbar disc

herniation. Int J Surg. 2016 Mar;27:8-16. doi:

10.1016/j.ijsu.2016.01.034. PMID: 26805569.

Exclusion Code: X1.

90. Li XC, Zhong CF, Deng GB, et al. Full-endoscopic

procedures versus traditional discectomy surgery for

discectomy: a systematic review and meta-analysis of

current global clinical trials. Pain Physician. 2016

Mar;19(3):103-18. PMID: 27008284. Exclusion

Code: X1.

91. Linzer P, Filip M, Samal F, et al. Comparison of

Biochemical Markers of Muscle Damage and

Inflammatory Response Between the Open

Discectomy, Microsurgical Discectomy, and

Microsurgical Discectomy Using Tubular Retractor. J



Neurol Surg A Cent Eur Neurosurg. 2015

Sep;76(5):384-91. doi: 10.1055/s-0034-1393929.


92. Livesey J, Sundaram S, Foster L, et al. Laser

discectomy versus lumbar epidural steroid injection:

a randomised comparative study of two treatments

for sciatica The Journal of Bone and Joint Surgery-

british Volume. 2000;82:74. Exclusion Code: X11.

93. Lurie JD, Faucett SC, Hanscom B, et al. Lumbar

discectomy outcomes vary by herniation level in the

Spine Patient Outcomes Research Trial. J Bone Joint

Surg Am. 2008 Sep;90(9):1811-9. doi:

10.2106/jbjs.g.00913. PMID: 18762639. Exclusion

Code: X8.

94. Lurie JD, Tosteson TD, Tosteson AN, et al. Surgical

versus nonoperative treatment for lumbar disc

herniation: eight-year results for the spine patient

outcomes research trial. Spine (Phila Pa 1976). 2014

Jan 01;39(1):3-16. doi:

10.1097/brs.0000000000000088. PMID: 24153171.

Exclusion Code: X9.

95. Machado GC, Witzleb AJ, Fritsch C, et al. Patients

with sciatica still experience pain and disability 5

years after surgery: A systematic review with meta-

analysis of cohort studies. Eur J Pain. 2016

Nov;20(10):1700-9. doi: 10.1002/ejp.893. PMID:


96. Macnab I. Negative Disc Exploration: an analysis of

the causes of nerve-root involvement in sixty-eight

patients. JBJS. 1971;53(5):891-903. PMID:

00004623-197153050-00004. Exclusion Code: X8.

97. Manchikanti L, Abdi S, Atluri S, et al. An update of

comprehensive evidence-based guidelines for

interventional techniques in chronic spinal pain. Part

II: guidance and recommendations. Pain Physician.

2013 Apr;16(2 Suppl):S49-283. PMID: 23615883.

Exclusion Code: X1.

98. Manchikanti L, Derby R, Benyamin RM, et al. A

systematic review of mechanical lumbar disc

decompression with nucleoplasty. Pain Physician.

2009 May-Jun;12(3):561-72. PMID: 19461823.

Exclusion Code: X9.

99. Manchikanti L, Falco FJ, Benyamin RM, et al. An

update of the systematic assessment of mechanical

lumbar disc decompression with nucleoplasty. Pain

Physician. 2013 Apr;16(2 Suppl):Se25-54. PMID:


100. Manchikanti L, Singh V, Calodney AK, et al.

Percutaneous lumbar mechanical disc decompression

utilizing Dekompressor(R): an update of current

evidence. Pain Physician. 2013 Apr;16(2 Suppl):Se1-


101. Manchikanti L, Singh V, Falco FJ, et al. An updated

review of automated percutaneous mechanical

lumbar discectomy for the contained herniated

lumbar disc. Pain Physician. 2013 Apr;16(2

Suppl):Se151-84. PMID: 23615890. Exclusion Code:

X8.

102. Marin FZ. CAM versus nucleoplasty. Acta Neurochir

Suppl. 2005;92:111-4. PMID: 15830980. Exclusion

Code: X6.

103. Martin-Laez R, Martinez-Agueros JA, Suarez-

Fernandez D, et al. Complications of endoscopic

microdiscectomy using the EASYGO! system: is

there any difference with conventional discectomy

during the learning-curve period? Acta Neurochir

(Wien). 2012 Jun;154(6):1023-32. doi:

10.1007/s00701-012-1321-5. PMID: 22446750.

Exclusion Code: X8.

104. Mazanec D, Okereke L. Interpreting the Spine Patient

Outcomes Research Trial. Medical vs surgical

treatment of lumbar disk herniation: implications for

future trials. Cleve Clin J Med. 2007 Aug;74(8):577-


105. McCormick ZL, Slipman C, Kotcharian A, et al.

Percutaneous Lumbar Disc Decompression Using the

Dekompressor: A Prospective Long-Term Outcome

Study. Pain Medicine. 2016;17(6):1023-30.

Exclusion Code: X8.

106. McGirt MJ, Ambrossi GL, Datoo G, et al. Recurrent

disc herniation and long-term back pain after primary

lumbar discectomy: review of outcomes reported for

limited versus aggressive disc removal.

Neurosurgery. 2009 Feb;64(2):338-44; discussion 44-

5. doi: 10.1227/01.neu.0000337574.58662.e2. PMID:


107. McGuire KJ, Khaleel MA, Rihn JA, et al. The effect

of high obesity on outcomes of treatment for lumbar

spinal conditions: subgroup analysis of the spine

patient outcomes research trial. Spine (Phila Pa

1976). 2014 Nov 01;39(23):1975-80. doi:

10.1097/brs.0000000000000577. PMID: 25365713.

Exclusion Code: X4.

108. Muralikuttan KP, Hamilton A, Kernohan WG, et al.

A prospective randomized trial of chemonucleolysis

and conventional disc surgery in single level lumbar

disc herniation. Spine (Phila Pa 1976). 1992

Apr;17(4):381-7. PMID: 1579871. Exclusion Code:

X6.

109. National Guideline Centre. National Institute for

Health and Care Excellence: Clinical Guidelines.

Low Back Pain and Sciatica in Over 16s: Assessment

and Management. London: National Institute for

Health and Care Excellence; 2016. Exclusion Code:

X1.

110. National Institute for Health and Care Excellence.

Automated percutaneous mechanical lumbar

discectomy. United Kingdom: National Institute for

Health and Care Excellence; 2005.

https://www.nice.org.uk/guidance/IPG141. Accessed

on November 8 2017. Exclusion Code: X1.




Percutaneous transforaminal endoscopic lumbar

discectomy for sciatica United Kingdom: National

Institute for Health and Care Excellence; 2016.

https://www.nice.org.uk/guidance/ipg556/resources/p

ercutaneous-transforaminal-endoscopic-lumbar-

discectomy-for-sciatica-pdf-1899871990618309.

Accessed on April 2017. Exclusion Code: X1.


Percutaneous interlaminar endoscopic lumbar

discectomy for sciatica United Kingdom: National

Institute for Health and Care Excellence; 2016.

https://www.nice.org.uk/guidance/ipg555. Accessed

on November 8 2017. Exclusion Code: X1.

113. Nellensteijn J, Ostelo R, Bartels R, et al.

Transforaminal endoscopic surgery for symptomatic

lumbar disc herniations: a systematic review of the

literature. Eur Spine J. 2010 Feb;19(2):181-204. doi:

10.1007/s00586-009-1155-x. PMID: 19756781.

Exclusion Code: X1.

114. Nikoobakht M, Yekanineajd MS, Pakpour AH, et al.

Plasma disc decompression compared to

physiotherapy for symptomatic contained lumbar disc

herniation: A prospective randomized controlled trial.

Neurol Neurochir Pol. 2016;50(1):24-30. doi:

10.1016/j.pjnns.2015.11.001. Epub 2015 Nov 28.


115. North American Spine Society. Clinical Guidelines

for Diagnosis and Treatment of Lumbar Disc

Herniation with Radiculopathy. Burr Ridge, IL:

Society NAS; 2012 2012.

https://www.spine.org/Portals/0/Documents/Research

ClinicalCare/Guidelines/LumbarDiscHerniation.pdf

Exclusion Code: X1.

116. Ohtori S, Koshi T, Yamashita M, et al. Surgical

versus nonsurgical treatment of selected patients with

discogenic low back pain: a small-sized randomized

trial. Spine (Phila Pa 1976). 2011 Mar 01;36(5):347-

54. doi: 10.1097/BRS.0b013e3181d0c944. PMID:


117. Ong D, Chua NH, Vissers K. Percutaneous Disc

Decompression for Lumbar Radicular Pain: A

Review Article. Pain Pract. 2016 Jan;16(1):111-26.

doi: 10.1111/papr.12250. PMID: 25354274.

Exclusion Code: X2.

118. Osterman H, Seitsalo S, Karppinen J, et al.

Effectiveness of microdiscectomy for lumbar disc

herniation: a randomized controlled trial with 2 years

of follow-up. Spine (Phila Pa 1976). 2006 Oct

01;31(21):2409-14. doi:

10.1097/01.brs.0000239178.08796.52. PMID:


119. Six-year outcomes of lumbar microdiscectomy:

results of a randomised controlled trial. Proceedings

of the International Society for the Study of the

Lumbar Spine; 2005. Exclusion Code: X11.

120. Pan L, Zhang P, Yin Q. Comparison of tissue

damages caused by endoscopic lumbar discectomy

and traditional lumbar discectomy: a randomised

controlled trial. Int J Surg. 2014;12(5):534-7. doi:

10.1016/j.ijsu.2014.02.015. PMID: 24583364.

Exclusion Code: X3.

121. Pan Z, Ha Y, Yi S, et al. Efficacy of Transforaminal

Endoscopic Spine System (TESSYS) Technique in

Treating Lumbar Disc Herniation. Med Sci Monit.

2016 Feb 18;22:530-9. PMID: 26887645. Exclusion

Code: X3.

122. Pauza KJ, Howell S, Dreyfuss P, et al. A randomized,

placebo-controlled trial of intradiscal electrothermal

therapy for the treatment of discogenic low back

pain. Spine J. 2004 Jan-Feb;4(1):27-35. PMID:


123. Payer M. "Minimally invasive" lumbar spine surgery:

a critical review. Acta Neurochir (Wien). 2011

Jul;153(7):1455-9. doi: 10.1007/s00701-011-1023-4.


124. Pearson A, Lurie J, Tosteson T, et al. Who should

have surgery for an intervertebral disc herniation?

Comparative effectiveness evidence from the spine

patient outcomes research trial. Spine (Phila Pa

1976). 2012 Jan 15;37(2):140-9. doi:

10.1097/BRS.0b013e3182276b2b. PMID: 21681140.

Exclusion Code: X8.

125. Pearson AM, Blood EA, Frymoyer JW, et al. SPORT

lumbar intervertebral disk herniation and back pain:

does treatment, location, or morphology matter?

Spine (Phila Pa 1976). 2008 Feb 15;33(4):428-35.

doi: 10.1097/BRS.0b013e31816469de. PMID:


126. Peul WC, van Houwelingen HC, van den Hout WB,

et al. Surgery versus prolonged conservative

treatment for sciatica N Engl J Med. 2007 May

31;356(22):2245-56. doi: 10.1056/NEJMoa064039.


127. Phan K, Xu J, Schultz K, et al. Full-endoscopic

versus micro-endoscopic and open discectomy: A

systematic review and meta-analysis of outcomes and

complications. Clin Neurol Neurosurg. 2017

Mar;154:1-12. doi: 10.1016/j.clineuro.2017.01.003.


128. Postacchini F, Lami R, Massobrio M.

Chemonucleolysis versus surgery in lumbar disc

herniations: correlation of the results to preoperative

clinical pattern and size of the herniation. Spine

(Phila Pa 1976). 1987 Mar;12(2):87-96. PMID:


129. Ran J, Hu Y, Zheng Z, et al. Comparison of

discectomy versus sequestrectomy in lumbar disc

herniation: a meta-analysis of comparative studies.

PLoS One. 2015;10(3):e0121816. doi:

10.1371/journal.pone.0121816. PMID: 25815514.

Exclusion Code: X1.



130. Rasouli MR, Rahimi-Movaghar V, Shokraneh F, et

al. Minimally invasive discectomy versus

microdiscectomy/open discectomy for symptomatic

lumbar disc herniation. Cochrane Database of

Systematic Reviews: John Wiley & Sons, Ltd; 2014.

Exclusion Code: X1.

131. Rasouli MR, Rahimi-Movaghar V, Shokraneh F, et

al. Minimally invasive discectomy versus

microdiscectomy/open discectomy for symptomatic

lumbar disc herniation. Cochrane Database Syst Rev.

2014 Sep 04(9):Cd010328. doi:

10.1002/14651858.CD010328.pub2. PMID:


132. Revel M, Payan C, Vallee C, et al. Automated

percutaneous lumbar discectomy versus

chemonucleolysis in the treatment of sciatica A

randomized multicenter trial. Spine (Phila Pa 1976).

1993 Jan;18(1):1-7. PMID: 8434309. Exclusion

Code: X6.

133. Reverberi C, Bottoli MG, Pennini M, et al. Disc

coablation and epidural injection of steroids: a

comparison of strategies in the treatment of

mechanical spinal discogenic pain. Acta Neurochir

Suppl. 2005;92:127-8. PMID: 15830983. Exclusion

Code: X8.

134. Righesso O, Falavigna A, Avanzi O. Comparison of

open discectomy with microendoscopic discectomy

in lumbar disc herniations: results of a randomized

controlled trial. Neurosurgery. 2007 Sep;61(3):545-9;

discussion 9. doi:

10.1227/01.neu.0000290901.00320.f5. PMID:


135. Rihn JA, Hilibrand AS, Radcliff K, et al. Duration of

symptoms resulting from lumbar disc herniation:

effect on treatment outcomes: analysis of the Spine

Patient Outcomes Research Trial (SPORT). J Bone

Joint Surg Am. 2011 Oct 19;93(20):1906-14. doi:

10.2106/jbjs.j.00878. PMID: 22012528. Exclusion

Code: X8.

136. Rowe SM ME-S, Song EK, Lee KB, Seo HY, Chung

JY. A prospective randomized study of percutaneous

endosc opicand microscopic lumbar disc ectomy.

Proceedings of theAmerican Academy of

Orthopaedic Surgeon; 2004. Exclusion Code: X11.

137. Ruan W, Feng F, Liu Z, et al. Comparison of

percutaneous endoscopic lumbar discectomy versus

open lumbar microdiscectomy for lumbar disc

herniation: A meta-analysis. Int J Surg. 2016

Jul;31:86-92. doi: 10.1016/j.ijsu.2016.05.061. PMID:


138. Ruetten S, Komp M, Merk H, et al. Surgical

treatment for lumbar lateral recess stenosis with the

full-endoscopic interlaminar approach versus

conventional microsurgical technique: a prospective,

randomized, controlled study. J Neurosurg Spine.

2009 May;10(5):476-85. doi: 10.3171/2008.7.17634.


139. S. A-F. Study of percutaneous intradiscal

nucleoplasty efficacy. Exclusion Code: X10.

140. Schenk B. Percutaneous laser disc decompression is

more cost-effective than conventional surgical

treatment for lumbar disc herniation and allows faster

patient rehabilitation, while long-term results are

comparable. ISRCTNR25884790. 2006. Exclusion

Code: X10.

141. Schenk B, Brouwer PA, Peul WC, et al. Percutaneous

laser disk decompression: a review of the literature.

AJNR Am J Neuroradiol. 2006 Jan;27(1):232-5.


142. Schick U, Dohnert J, Richter A, et al.

Microendoscopic lumbar discectomy versus open

surgery: an intraoperative EMG study. Eur Spine J.

2002 Feb;11(1):20-6. PMID: 11931060. Exclusion

Code: X7.

143. Schizas C, Tsiridis E, Saksena J. Microendoscopic

discectomy compared with standard microsurgical

discectomy for treatment of uncontained or large

contained disc herniations. Neurosurgery. 2005

Oct;57(4 Suppl):357-60; discussion -60. PMID:


144. Schoenfeld AJ, Bono CM. Does surgical timing

influence functional recovery after lumbar

discectomy? A systematic review. Clin Orthop Relat

Res. 2015 Jun;473(6):1963-70. doi: 10.1007/s11999-

014-3505-1. PMID: 24526298. Exclusion Code: X7.

145. Shamji MF, Bains I, Yong E, et al. Treatment of

herniated lumbar disk by sequestrectomy or

conventional diskectomy. World Neurosurg. 2014

Nov;82(5):879-83. doi: 10.1016/j.wneu.2013.02.066.


146. Shih P, Wong AP, Smith TR, et al. Complications of

open compared to minimally invasive lumbar spine

decompression. J Clin Neurosci. 2011

Oct;18(10):1360-4. doi: 10.1016/j.jocn.2011.02.022.


147. Shin DA, Kim KN, Shin HC, et al. The efficacy of

microendoscopic discectomy in reducing iatrogenic

muscle injury. J Neurosurg Spine. 2008 Jan;8(1):39-

43. doi: 10.3171/spi-08/01/039. PMID: 18173345.

Exclusion Code: X7.

148. Shriver MF, Xie JJ, Tye EY, et al. Lumbar

microdiscectomy complication rates: a systematic

review and meta-analysis. Neurosurg Focus. 2015

Oct;39(4):E6. doi: 10.3171/2015.7.focus15281.


149. Singh V, Benyamin RM, Datta S, et al. Systematic

review of percutaneous lumbar mechanical disc

decompression utilizing Dekompressor. Pain

Physician. 2009 May-Jun;12(3):589-99. PMID:




150. Singh V, Manchikanti L, Benyamin RM, et al.

Percutaneous lumbar laser disc decompression: a

systematic review of current evidence. Pain

Physician. 2009 May-Jun;12(3):573-88. PMID:


151. Singh V, Manchikanti L, Calodney AK, et al.

Percutaneous lumbar laser disc decompression: an

update of current evidence. Pain Physician. 2013

Apr;16(2 Suppl):Se229-60. PMID: 23615885.

Exclusion Code: X8.

152. Smith N, Masters J, Jensen C, et al. Systematic

review of microendoscopic discectomy for lumbar

disc herniation. Eur Spine J. 2013 Nov;22(11):2458-

65. doi: 10.1007/s00586-013-2848-8. PMID:


153. Steffen R. A prospective comparative study of

chemonucleolysis and laser discectomy. Orthop

Trans. 1996;20:388. Exclusion Code: X6.

154. Steffen R. Chemonucleolysis versus laser disc

decompression-a prospective randomised trial. J

Bone Joint Surg, B. 1997;79:247. Exclusion Code:

X11.

155. Steffen R. Laser discotomy/chemonucleolysis ini the

therapy of displaced lumbal intervertebral disc—

prospective randomised comparison. Hefte zur der

Unfallchirurg. 1999;271:215-21. Exclusion Code:

X12.

156. Stula D. [Chemonucleolysis with chymopapain in

lumbar disk hernia Randomized comparative study

with operated patients]. Neurochirurgia (Stuttg). 1990

Sep;33(5):169-72. doi: 10.1055/s-2008-1053577.


157. Thomas KC, Fisher CG, Boyd M, et al. Outcome

evaluation of surgical and nonsurgical management

of lumbar disc protrusion causing radiculopathy.

Spine (Phila Pa 1976). 2007 Jun 01;32(13):1414-22.

doi: 10.1097/BRS.0b013e318060a5d1. PMID:


158. Thome C, Borm W, Meyer F. Degenerative lumbar

spinal stenosis: current strategies in diagnosis and

treatment. Dtsch Arztebl Int. 2008 May;105(20):373-

9. doi: 10.3238/arztebl.2008.0373. PMID: 19626175.

Exclusion Code: X4.

159. Tohmeh AG, Smith WD. Lumbar total disc

replacement by less invasive lateral approach: a

report of results from two centers in the US IDE

clinical trial of the XL TDR(R) device. Eur Spine J.

2015 Apr;24 Suppl 3:331-8. doi: 10.1007/s00586-

015-3843-z. PMID: 25794698. Exclusion Code: X5.

160. Tosteson AN, Skinner JS, Tosteson TD, et al. The

cost effectiveness of surgical versus nonoperative

treatment for lumbar disc herniation over two years:

evidence from the Spine Patient Outcomes Research

Trial (SPORT). Spine (Phila Pa 1976). 2008 Sep

01;33(19):2108-15. PMID: 18777603. Exclusion

Code: X9.

161. Toyoda H, Nakamura H, Konishi S, et al. Clinical

outcome of microsurgical bilateral decompression via

unilateral approach for lumbar canal stenosis:

minimum five-year follow-up. Spine (Phila Pa 1976).

2011 Mar 01;36(5):410-5. doi:

10.1097/BRS.0b013e3181d25829. PMID: 20847714.

Exclusion Code: X4.

162. Toyone T, Tanaka T, Kato D, et al. Low-back pain

following surgery for lumbar disc herniation. A

prospective study. J Bone Joint Surg Am. 2004

May;86-a(5):893-6. PMID: 15118028. Exclusion

Code: X8.

163. Tureyen K. One-level one-sided lumbar disc surgery

with and without microscopic assistance: 1-year

outcome in 114 consecutive patients. J Neurosurg.

2003 Oct;99(3 Suppl):247-50. PMID: 14563140.

Exclusion Code: X3.

164. Urrutia G, Kovacs F, Nishishinya MB, et al.

Percutaneous thermocoagulation intradiscal

techniques for discogenic low back pain. Spine (Phila

Pa 1976). 2007 May 01;32(10):1146-54. doi:

10.1097/01.brs.0000261492.55121.93. PMID:


165. Vaccaro AR, Fehlings MG. The applicability of

clinical equipoise and sham surgery in patients with

symptomatic lumbar radiculopathy due to a herniated

disc: the SPORT trial. LWW; 2007. Exclusion Code:

X2.

166. van Alphen HA, Braakman R, Berfelo MW, et al.

Chemonucleolysis or discectomy? Results of a

randomized multicentre trial in patients with a

herniated lumbar intervertebral disc (a preliminary

report). Acta Neurochir Suppl (Wien). 1988;43:35-8.


167. van Alphen HA, Braakman R, Bezemer PD, et al.

Chemonucleolysis versus discectomy: a randomized

multicenter trial. J Neurosurg. 1989 Jun;70(6):869-

75. doi: 10.3171/jns.1989.70.6.0869. PMID:


168. Veresciagina K, Spakauskas B, Ambrozaitis KV.

Clinical outcomes of patients with lumbar disc

herniation, selected for one-level open-discectomy

and microdiscectomy. Eur Spine J. 2010

Sep;19(9):1450-8. doi: 10.1007/s00586-010-1431-9.


169. A randomised prospective study comparing laser

decompression and epidural steroid injection in

alleviating radicular pain secondary to prolapsed

lumbar discs. Proceedings of the International

Society for Study of theLumbar Spine; 1998.


170. Wang H, Lü F, Jiang J, et al. Minimally invasive

lumbar interbody fusion via MAST Quadrant

retractor versus open surgery: a prospective

randomized clinical trial. Chinese medical journal;

2011. p. 3868-74. Exclusion Code: X5.



171. Wang XS, Sun RF, Ji Q, et al. A meta-analysis of

interlaminar minimally invasive discectomy

compared to conventional microdiscectomy for

lumbar disk herniation. Clin Neurol Neurosurg. 2014

Dec;127:149-57. doi:

10.1016/j.clineuro.2014.10.001. PMID: 25459263.

Exclusion Code: X1.

172. Watters WC, 3rd, McGirt MJ. An evidence-based

review of the literature on the consequences of

conservative versus aggressive discectomy for the

treatment of primary disc herniation with

radiculopathy. Spine J. 2009 Mar;9(3):240-57. doi:

10.1016/j.spinee.2008.08.005. PMID: 18809356.

Exclusion Code: X1.

173. Watters WC, 3rd, Mirkovic S, Boss J. Treatment of

the isolated lumbar intervertebral disc herniation:

microdiscectomy versus chemonucleolysis. Spine

(Phila Pa 1976). 1988 Mar;13(3):360-2. PMID:


174. Weinstein JN, Lurie JD, Tosteson TD, et al. Surgical

vs nonoperative treatment for lumbar disk herniation:

the Spine Patient Outcomes Research Trial (SPORT)

observational cohort. Jama 2006 Nov

22;296(20):2451-9. doi: 10.1001/jama.296.20.2451.


175. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical

versus nonoperative treatment for lumbar spinal

stenosis four-year results of the Spine Patient

Outcomes Research Trial. Spine (Phila Pa 1976).

2010 Jun 15;35(14):1329-38. doi:

10.1097/BRS.0b013e3181e0f04d. PMID: 20453723.

Exclusion Code: X8.

176. Wu S, Li X, Lin C, et al. CT-guided nucleoplasty

with radiofrequency energy for the treatment of

lumbar disk herniation. J Spinal Disord Tech. 2015

Feb;28(1):E9-16. doi:

10.1097/bsd.0000000000000132. PMID: 25023711.

Exclusion Code: X3.

177. Wu Z, Wei LX, Li J, et al. Percutaneous treatment of

non-contained lumbar disc herniation by injection of

oxygen-ozone combined with collagenase. Eur J

Radiol. 2009 Dec;72(3):499-504. doi:

10.1016/j.ejrad.2008.07.029. PMID: 18778905.

Exclusion Code: X3.

178. Xu HH, Wang XQ, Zhang YL, et al. Efficacy

evaluation for the treatment of unstable lumbar disc

herniation by traditional and modified lamina

osteotomy replantation. Chin Med J (Engl).

2013;126(15):2840-5. PMID: 23924453. Exclusion

Code: X5.

179. Yao Y, Zhang H, Wu J, et al. Comparison of Three

Minimally Invasive Spine Surgery Methods for

Revision Surgery for Recurrent Herniation After

Percutaneous Endoscopic Lumbar Discectomy.

World Neurosurg. 2017 Apr;100:641-7.e1. doi:

10.1016/j.wneu.2017.01.089. PMID: 28153616.

Exclusion Code: X3.

180. Ying J, Huang K, Zhu M, et al. The Effect and

Feasibility Study of Transforaminal Percutaneous

Endoscopic Lumbar Discectomy Via Superior Border

of Inferior Pedicle Approach for Down-Migrated

Intracanal Disc Herniations. Medicine (Baltimore).

2016 Feb;95(8):e2899. doi:

10.1097/md.0000000000002899. PMID: 26937930.

Exclusion Code: X3.

181. Yoshihara H, Chatterjee D, Paulino CB, et al.

Revision Surgery for "Real" Recurrent Lumbar Disk

Herniation: A Systematic Review. Clin Spine Surg.

2016 Apr;29(3):111-8. doi:

10.1097/bsd.0000000000000365. PMID: 27002374.

Exclusion Code: X7.

182. Zhang L, Ding XL, Zhao XL, et al. Fluoroscopy-

guided Bipolar Radiofrequency Thermocoagulation

Treatment for Discogenic Low Back Pain. Chin Med

J (Engl). 2016 Oct 05;129(19):2313-8. doi:

10.4103/0366-6999.190682. PMID: 27647190.

Exclusion Code: X4.

183. Zhang Q, Qian J, Zhu Y. Meta-Analysis on

Microdiscectomy and Sequestrectomy for Lumbar

Disc Herniation. J Invest Surg. 2015;28(4):225-9.


184. Zhu H, Zhou XZ, Cheng MH, et al. The efficacy of

coblation nucleoplasty for protrusion of lumbar

intervertebral disc at a two-year follow-up. Int

Orthop. 2011 Nov;35(11):1677-82. doi:

10.1007/s00264-010-1196-0. PMID: 21240606.

Exclusion Code: X8.


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page E-1

Appendix E. Individual Study Risk of Bias Assessments

Table E-1. Risk of bias ratings-Overall rating and randomization process

Author (Year)

Overall Bias Randomization Process Bias

Overall Quality Rating

Overall Rationale for Quality Rating

Was the allocation sequence random?

Was allocation sequence concealed until participants were recruited and assigned to interventions?

Were there baseline imbalances that suggest a problem with the randomization process?

Bias arising from randomization or selection? Comments

Arts (2009)38 Arts (2011)43

Low This risk of bias assessment includes both 1y and 2y results.

Yes Yes No Low

Brouwer (2015)37 Brouwer (2017)97

Some concerns

Some concerns for bias because of lack of blinding of study participants, who also served as outcome assesors.

Yes Yes No Low


Some concerns for bias because of no information on randomization methods, patients were not blinded, and unclear whether outcome assessors were fully blinded or if patient self-reported outcomes were used.

No information No information No information Some concerns No information provided at all about how participants were randomized and how allocation was concealed, and the study did not provide any comparison of baseline characteristics between groups to assess adequancy of randomization.

(continued)



Table E-1. Risk of bias ratings-Overall rating and randomization process (continued)

Author (Year)









High High risk of bias concerns due to inadequate method of randomization. Some concerns for risk of bias due to lack of participant blinding and use of patient self-reported outcomes. Low risk of bias for some safety outcomes.

No information Probably no No information High Evenly numbered referral patients were assigned to surgery and odd-numbered referral patients were assigned to conservative care. This is an inadequate method of randomization. Only gender and age are reported at basleline; there were differences in baseline pain between the groups, but these differences were small.

(continued)




Author (Year)









Some concerns for bias due to no information about baseline balance between groups, lack of participant blinding, including outcome assessment with patient self-reported outcomes.

Probably yes No information No information Some concerns Authors report no significant group difference for preop parameters but don't provide data.

Gerszten (2010)39 Some concerns for outcomes at 6w; high for outcomes at 12w or later.

Some concerns for bias because patients were not blinded and self-reported outcomes used at all time points. High risk of bias for outcomes reported at 12w and later because of second procedures provided to participants in both groups, and high attrition at 12w and later. Low risk of bias for some safety outcomes.

Probably yes Yes Probably no Low Mean duration of radicular pain was double in the epidural injection group, but the sample size was small and range was large.

(continued)




Author (Year)








Haines (2002)40 High High risk of bias because of randomization process and high attrition in both groups, also some concerns for bias because participants and clinician outcome assessors not blinded.

No information No information Probably yes High No information about randomization process and allocation concealment combined with imbalances in randomization (21 randomized to APLD 13 randomized to discectomy); average age was 7 years less in the discectomy group, but this difference was not statistically significant.


Some concerns because of randomization process/allocation concealment unconventional and no baseline characteristic to assess balance and use of patient self-reported outcomes for unblinded participants.

Probably yes Probably yes No information Some concerns Not a conventional approach to generating random sequence, and no table of baseline characteristics to assess whether groups were balanced at baseline.

(continued)




Author (Year)









Some concerns for bias because of use of patient-self reported outcomes with participants that were not blinded.

No information Probably yes Probably no Low


Some concerns for bias because of randomication/allocation concealment process and use of patient-reported outcomes when participants were not blinded.

No information No information No information Some concerns

Mayer (1993)32 High Some concerns for bias over randomization process, deviations from interventions, and use of patient-reported outcomes among participants that were not blinded. Because of some concerns in 3 of the 5 domains, this increases the level of concern overall to high.

No information No information Probably no Some concerns

(continued)




Author (Year)









Some concerns for risk of bias because study used patient self-reported outcomes and participants were not blinded to treatment allocation. Note: this risk of bias rating only applies to outcomes up to 12 weeks for the intent-to-treat analysis. Crossovers were allowed after 12 weeks and raise the risk of bias to high for outcomes after 12 weeks. Low risk of bias for some safety outcomes.

Yes Probably yes No Low

(continued)




Author (Year)








North (2005)47 High High risk of bias because of extensive deviations from intended interventions as a result of crossovers, also differential attrition. Some concerns for patient self-reported outcomes when participants were not blinded.

Yes Yes No information Low

Osterman (2003)31 High High risk of bias because of extensive differential crossover from physiotherapy to surgery, and differential co-interventions related to the outcome. Also, some concerns for patient-reported outcomes when participants were not blinded. Low risk of bias for some safety outcomes.

Yes Yes No Low

(continued)




Author (Year)








Peul (2007)30 Peul (2008)95 Lequin (2013)96

High Moderate to extensive deviations from intended interventions as a result of crossovers, which are likely non-random. This most likely biases the effect toward the null. Some concerns from use of patient-reported outcomes when participants were not blinded. Low risk of bias for some safety outcomes.

Yes Yes No Low

Ruetten (2008)29 High Randomization and allocation concealment process was inadequate and thus, subjects study to high risk of bias. Some concerns for bias because of use of self-reported outcomes among participants that were not blinded.

Probably no Probably no No information High Randomization was peformed by alternate assignment of participants. No baseline characteristics are provided to assess balance between groups.

(continued)




Author (Year)








Ruetten (2009)48 High Randomization and allocation concealment process was inadequate and thus, subjects study to high risk of bias. Some concerns for bias because of use of self-reported outcomes among participants that were not blinded.

Probably no Probably no No information High Randomization was peformed by alternate assignment of participants. No baseline characteristics are provided to assess balance between groups.

Ryang (2008)28 Gempt (2013)98

Some concerns

Some concerns for bias in three of the five domains elevates this to level of concern for bias to high. Some concerns over randomization process, information about missing data, and use of patient-reported outcomes among participants that were not blinded.

No information No information Probably no Some concerns Method of randomization and allocation concealment not reported; only age and gender reported at baseline to assess baseline comparability.

(continued)




Author (Year)








Sasaoka (2006)24 High Some concerns for bias across 4 of the 5 domains increase the overall risk of bias to high. Some concerns for bias in randomization, deviations from interventions, missing data and measurement of outcome.

No information No information No information Some concerns No details on methods of randomization and allocation concealment, only sex and age, and location of herniation was reported by group so unable to tell if groups were balanced at baseline.


Some concerns for bias because patient-reported outcomes were used and participants were not blinded to treatment allocation.

Probably yes No information No Low No details on allocation concealment.

Thome (2005)26 Barth (2008)99

Some concerns

Some concerns for bias because patient-reported outcomes were used and participants were likely not blinded to treatment allocation.

Yes Probably yes No Low Randomization was reported as "concealed" but no specific details as to how it was concealed were provided.

(continued)




Author (Year)









Some concerns for bias because methods of randomization and allocation concealment were not reported and not enough information was available to judge baseline balance among groups. Also some concerns for bias because patient-reported outcomes were used and participants were not blinded to treatment allocation.

No information No information No information Some concerns Method of randomization and allocation concealment not reported. Very little information was provided except sex, age, length of time off work prior to surgery, and workloads to assess balance between groups. The discectomy group was biased toward heavier workloads, but sample size is small so difference may not be significant.

(continued)




Author (Year)








Weber (1983)5 High High risk of bias because extensive crossovers, also some concerns because patient-reported outcomes were used and participants were not blinded.

Probably yes Probably yes No information Low Age and gender are the only characteristics reported

Weinstein (2006) 21 Weinstein (2008)45 Lurie Jon (2014)44

High High risk of bias because of extensive deviations from intended interventions as a result of crossovers, which are likely non-random. This most likely biases the effect toward the null. Some concerns for bias because of patient-reported outcomes when participants were not blinded. Low risk of bias for some safety outcomes.

Yes Probably yes No Low



Table E-2. Risk of bias —deviations from intended interventions

Author (Year)

Were the participants aware of their assigned intervention during the trial?

Were carers and trial personnel aware of participants' assigned intervention during the trials?

Were there deviations from the intended intervention beyond what would be expected in usual practice?

Were these deviations from intended intervention unbalanced between groups and likely to have affected the outcome?

Were any participants analyzed in a group different from the one to which they were assigned?

Was there potential for a substantial impact of analyzing participants in the wrong group?

Bias arising from deviations from intended interventions? Comments

Arts (2009)38 Arts (2011)43

No Yes No NA No NA Low Not feasible to blind surgeons and other caregivers to the intervention assignment.


Yes Yes No NA No NA Low

Chatterjee (1995)36

Yes Yes No NA No NA Low It was not reported whether participants were blinded, but the study treatments are different enough that it would be obvious.


Yes Probably no Probably no NA Probably no NA Low

Franke (2009)34

No information No information Probably no NA No NA Low



Table E-2. Risk of bias —deviations from intended interventions (continued)

Author (Year)








Gerszten (2010)39

Yes Yes Probably yes Yes Probably no NA Some concerns for 6w outcomes; High for outcomes at 12w or later.

5 patients were randomized but did not receive treatment (1 in surgical group, 4 in nonsurgical group). 30 of 40 patients randomized to epidural injections received both injections; 10 opted to not receive both injections. By 26w, 12 participants in surgery group received additional unspecfied second procedure. By 26w, 8 participants in the epidural injection received additional unspecfiied second procedures. These numbers are 9 and 5, respectively, at 12w.




Author (Year)








Haines (2002)40

No information Yes No NA No NA Low

Henriksen (1996)33

No Probably yes Probably no NA Probably no NA Low

Hermantin (1999)41

Yes Yes No NA No NA Low

Huang (2005)23

No information Yes No NA No NA Low

Mayer (1993)32

No information Yes Probably yes Probably yes No NA Some concerns

3 patients in percutaneous surgery group also received microdiscectomy as a second procedure.

McMorland (2010)22

Yes Yes Probably no NA No NA Low Pertains to outcomes at 12w




Author (Year)







Bias arising from deviations from intended interven-tions? Comments

North (2005)47

Yes Yes Yes Yes No information NA High 10 patients (4 in reoperation group, 6 in spinal cord stimulation group) were excluded postrandomization because of failure to receive authorization. 5 (21%) of those allocated to spinal cord stimulation also received reoperation. 14 (54%) of those allocated to reoperation also received spinal cord stimulation.




Author (Year)








Osterman (2003)31

Yes Yes Yes Yes No NA High 0% allocated to surgery received physiotherapy; 39% allocated to physiotherapy received surgery. In addition, 15 participants in the physiotherapy group and 8 participants in the surgery group reported receiving additional physical therapy outside of the study during followup.




Author (Year)








Peul (2007)30 Peul (2008)95 Lequin (2013)96

Yes Probably yes Yes Probably yes No NA High Moderate deviation in surgical group, 11.3% did not receive surgery; extensive deviation in nonsurgical group, 39% received surgery by 1 year and 44% received surgery by 2 years. These deviations make the groups more similar to each other and are very likely to bias results toward the null.

Ruetten (2008)29

Yes Probably yes Probably no NA Probably no NA Low The physicians performing the 2-year exams were not involved in the surgeries, but it is probably visually clear what surgery may have been performed.

Ruetten (2009)48

Yes Probably yes Probably no NA Probably no NA Low



Table E-2. Risk of bias for surgery—deviations from intended interventions (continued)

Author (Year)








Ryang (2008)28 Gempt (2013)98

No information Yes Probably no NA No NA Low

Sasaoka (2006)24

No information No information Probably no NA Probably no NA Some concerns

No information about blinding of participants or clinicians, no information about deviations from interventions though likely crossovers would have been reported if they occurred.

Teli (2010)27 Yes Probably yes Probably no NA No NA Low

Thome (2005)26 Barth (2008)99

No information No information No NA Probably no NA Low

Tullberg (1993)25

No information No information Probably no NA Yes Probably no Low No information about blinding of participants or clinicians, no information about deviations from interventions.



Table E-2. Risk of bias—deviations from intended interventions (continued)

Author (Year)








Weber (1983)5

Yes Probably yes Probably yes Probably yes No NA High Over a quarter of participants allocated to conservative management were referred for surgery within a year of randomization. Only 2 surgical patients did not have the operation as randomized. Authors report the data so an ITT analysis can be performed.




Author (Year)









Yes Yes Yes Yes No NA High Extensive deviation. Of those assigned to surgery, 50% received surgery by 3 months (60% by 2 years); of those assigned to no surgery, 30% received surgery by 3 months (45% by 2 years). These deviations make the groups more similar to each other and are very likely to bias results toward the null.



Table E-3. Risk of bias —missing outcome data

Author (Year)

Were outcome data available for all, or nearly all, participants randomized?

Are the proportions of missing outcome data and reasons for missing outcome data similar across intervention groups?

Is there evidence that results were robust to the presence of missing outcome data?

Bias arising from missing outcome data? Comments

Arts (2009)38 Arts (2011)43

Yes NA NA Low SG1: 94% had data available at 1y, 88% had data available at 2y. 166 were included in primary analysis. SG2: 93% had data available at 1y, 91% had data available at 2y. 159 were included in primary analysis.


Yes NA NA Low

Chatterjee (1995)36

Yes NA NA Low


Probably yes NA NA Low

Franke (2009)34 Yes NA NA Low

Gerszten (2010)39 No Probably yes Probably yes Some concerns at 6w, high for outcomes at 12w or later.

High attrition in both groups at 12w and 26w (surgery attrition 34%; epidural steroid group attrition 30%)

Haines (2002)40 No Probably no No information High Conflicting information about number of participants lost to followup; 24% attrition in APLD grup; 31% attrition in discectomy group; overa attrition at 52w was 44%.

Henriksen (1996)33 Yes NA NA Low

Hermantin (1999)41

Yes NA NA Low

Huang (2005)23 Yes NA NA Low

Mayer (1993)32 Yes NA NA Low

McMorland (2010)22

Yes NA NA Low



Table E-3. Risk of bias —missing outcome data (continued)

Author (Year)





North (2005)47 Probably no No No information High Loss of followup was exclusively in the spinal stimulation group. 82% of randomized patients provided short-term results; 75% of randomized provided long-term results. 98% of the randomized and treated patients provided short-term results; 90% of the rand

Osterman (2003)31 Yes NA NA Low

Peul (2007)30 Peul (2008)95 Lequin (2013)96

Yes NA NA Low

Ruetten (2008)29 Yes NA NA Low

Ruetten (2009)48 Yes NA NA Low

Ryang (2008)28 Gempt (2013)98

No information No information No information Some concerns No information provided as to how many participants contributed data at followup; no way to ascertain whether missing data were present.

Sasaoka (2006)24 No information No information No information Some concerns No information about how many participants contributed follow-up data.

Teli (2010)27 Yes NA NA Low

Thome (2005)26 Barth (2008)99

Yes NA NA Low



Table E-3. Risk of bias —missing outcome data (continued)

Author (Year)





Tullberg (1993)25 Yes NA NA Low 2 patients, 1 in each group, underwent reoperations for recurrence within a year of the first surgery and were not included among the overall study's results.

Weber (1983)5 Probably yes NA NA Low

Weinstein (2006)21 Weinstein (2008)45 Lurie Jon (2014)44

Yes NA NA Low 5.3% missing in surgery group; 6.2% missing in nonsurgery group.



Table E-4. Risk of bias —measurement of the outcome

Author (Year)

Were outcome assessors aware of the intervention received by study participants?

Was the assessment of the outcome likely to be influenced by knowledge of intervention received?

Bias arising from measurement of the outcome? Comments

Arts (2009)38 Arts (2011)43

Probably no NA Low Participants and observers were blinded to allocated treatment during the follow-up period.


Yes Probably yes Some concerns Many patient-reported outcomes used as appropriate and not feasible to blind participants to study intervention, though most participants had no stated preference for treatment suggesting that risk of bias is probably not high.

Chatterjee (1995)36

Probably no NA Some concerns Not enough information to judge were outcome assessors were truly blinded or if patient contributed towards outcomes.


Probably yes Probably yes Some concerns Uses some patient self-reported outcomes, and intervention was not blinded to participants.

Franke (2009)34 No information Probably yes Some concerns Since most outcomes are patient-reported and not clear that allocation was blinded, there is some concern for bias, though probably only small given the comparison in this study is between two surgical treatments.

Gerszten (2010)39 Yes Probably yes Some concerns Study uses patient-reported outcomes, thus some risk of bias is possible.

Haines (2002)40 No information Probably yes Some concerns Assessment matrix included both clinician and patient-reported outcomes; neither were blinded.

Henriksen (1996)33 Probably yes Probably yes Some concerns Use of patient self-reported outcomes, participants weren't blinded.

Hermantin (1999)41

Yes Probably yes Some concerns Study uses patient-reported outcomes, so some concern for bias as may be influenced by knowledge of treatment allocation.



Table E-4. Risk of bias —measurement of the outcome (continued)

Author (Year)




Huang (2005)23 No information Probably yes Some concerns Study used patient-reported outcomes, which may be influenced by knowledge of treatment allocation.

Mayer (1993)32 Probably yes Probably yes Some concerns Use of patient self-reported outcomes, participants weren't blinded.

McMorland (2010)22

No information Probably yes Some concerns Study used patient-reported outcomes, which have some concern for bias given that participants were not blinded to the treatment allocation.

North (2005)47 Yes Probably yes Some concerns Uses some patient self-reported outcomes, and treatment allocation was not blinded.

Osterman (2003)31 Yes Probably yes Some concerns Study used patient-reported outcomes, and there is some concern that these can be influenced by knowledge of treatment allocation. Study evaluated patient and surgeon's expectations of improvement immediately after randomization and showed higher expectations.

Peul (2007)30 Peul (2008)95 Lequin (2013)96

Yes Probably yes Some concerns Study used patient-reported outcomes, which are likely to be influenced by knowledge of treatment allocation since participants were not blinded.

Ruetten (2008)29 Probably yes Probably yes Some concerns Most outcome measures were patient self-report; participants were not blind to treatment allocation.

Ruetten (2009)48 Probably yes Probably yes Some concerns Most outcome measures were patient self-report; participants were not blind to treatment allocation.



Table E-4. Risk of bias —measurement of the outcome (continued)

Author (Year)




Ryang (2008)28 Gempt (2013)98

No information Probably yes Some concerns Use of patient-reported outcomes among participants,who were likely not blinded to treatment allocation.

Sasaoka (2006)24 No information No information Some concerns Use of patient-reported outcomes by participants who were likely not blinded to treatment allocation.

Teli (2010)27 Yes Probably yes Some concerns Patient self-reported outcomes; patients were not blind to treatment allocation.

Thome (2005)26 Barth (2008)99

No information Probably yes Some concerns Use of patient-reported outcomes among participants who were likely not blinded to treatment allocation.

Tullberg (1993)25 No information No information Some concerns Use of patient-reported outcomes by participants who were likely not blinded to treatment allocation.

Weber (1983)5 Yes Probably yes Some concerns Patient-reported outcomes were used among participants who were not blinded.


No information No information Some concerns Study used patient-reported outcomes, which are likely to be influenced by knowledge of treatment allocation as participants were not blinded.



Table E-5. Risk of bias —selection of the reported result

Author (Year)

Are the reported outcome data likely to have been selected on the basis of results from multiple outcome measurements within the outcome domain?

Are the reported outcome data likely to have been selected on the basis of results from multiple analyses of the data?

Bias arising from selection of reported results? Comments

Arts (2009)38 Arts (2011)43

No No Low


No No Low

Chatterjee (1995)36 No No Low

Erginousakis (2011)35 Probably no Probably no Low

Franke (2009)34 No No Low

Gerszten (2010)39 No No Low

Haines (2002)40 No No Low

Henriksen (1996)33 No No Low

Hermantin (1999)41 No No Low

Huang (2005)23 No No Low

Mayer (1993)32 No No Low

McMorland (2010)22 No No Low

North (2005)47 Probably no Probably no Low

Osterman (2003)31 No No Low

Peul (2007)30 Peul (2008)95 Lequin (2013)96

No No Low

Ruetten (2008)29 Probably no Probably no Low

Ruetten (2009)48 Probably no Probably no Low

Ryang (2008)28 Gempt (2013)98

No No Low

Sasaoka (2006)24 No No Low

Teli (2010)27 Probably no Probably no Low

Thome (2005)26 Barth (2008)99

No No Low

Tullberg (1993)25 Probably no Probably no Low

Weber (1983)5 Probably no Probably no Low


No No Low Authors also conducted an 'as treated' analysis in addition to the intent-to-treat analysis, but this risk of bias assessment is only focused on the intent-to-treat analysis.


Surgery for Lumbar Radiculopathy/Sciatica: Draft evidence report Page F-1

Appendix F. Meta-analyses

Figure F-1. Between-group differences in visual analogue scale for leg pain in randomized controlled trials comparing minimally invasive surgery with standard surgery at 12 weeks to 26 weeks, 52 weeks to 1.5 years, and 2 years.

Note: This figure only depicts the microendoscopic discectomy compared with microdiscectomy comparison reported by Teli et al.27 The VAS 10 cm scores reported by Thome et al.26 were converted to 100 mm for this analysis. Mean follow-up scores for Arts et al.38 and Brouwer et al.37 are unadjusted for baseline, so the AMDs reported here may differ from the AMDs reported by the study publications.



Figure F-2. Between-group differences in visual analogue scale for back pain in randomized controlled trials comparing minimally invasive surgery with standard surgery at 12 weeks to 26 weeks, 52 weeks to 1.5 years, and 2 years.

Note: This figure only depicts the microendoscopic discectomy compared with microdiscectomy comparison reported by Teli et al.27 The VAS 10 cm scores reported by Thome et

al.26 were converted to 100 mm for this analysis. Mean follow-up scores for Arts et al.38 and Brouwer et al.37 are unadjusted for baseline, so the AMDs reported here may differ

from the AMDs reported by the study publications.



Figure F-3. Between-group differences in SF-36 Bodily Pain subscale in randomized controlled trials comparing minimally invasive surgery with standard surgery at 12 weeks to 26 weeks.

Note: Mean follow-up scores for all studies are unadjusted for baseline, so the AMDs reported here may differ from the AMDs reported by the study publications.

Figure F-4. Between-group differences in SF-36 Physical Functioning subscale in randomized controlled trials comparing minimally invasive surgery with standard surgery at 12 weeks to 26 weeks.

Note: Mean follow-up scores for all studies are unadjusted for baseline, so the AMDs reported here may differ from the AMDs reported by the study publications.